Local economic development driven by gastronomy: the figures that actually move a territory

How much employment does gastronomy actually move in a territory?

Accommodation and food services sustain roughly 6.5% of total employment in Latin America and the Caribbean according to the ILO, with informality hovering near 70% in several countries of the bloc, and that mix explains why a local economic development program pays off here.

Consider the arithmetic facing a program officer: if two hundred establishments operate in your municipality and each employs between three and seven people, you are managing a labor mass no free trade zone will replicate within five years. The business structure reinforces the argument, since 95% of Colombia's gastronomic market consists of independent establishments according to Acodrés in 2024, owners who decide within the same week and without a corporate committee. The decision this figure triggers is simple and unpopular: stop counting workshop beneficiaries and start counting contracts with an active social security payroll record at twenty-four months. An owner formalizes when the contract costs less than turnover, and that comparison runs on numbers, not on development rhetoric.

Formality is not decreed: it becomes profitable or it never happens

With sectoral informality close to 70% across the region according to the ILO, any program offering only free training is giving away something the operator values less than this week's cash. I got this wrong for years, pushing costing courses when the real problem sat elsewhere: nobody knew what replacing a grill cook actually cost. A server who quits after four months forces you to repeat the hiring, the training and the first weeks of expediting mistakes, and that hidden cost competes head-on against the payroll burden the owner keeps dodging. Once the program puts both figures on the same sheet, formalization stops being a favor to the State and becomes a margin decision. Measure the share of formal payroll per establishment, never the municipal aggregate, because the aggregate hides the five large players that already complied. Surplus food in US foodservice reached USD 157 billion in 2024, equal to 14% of sector sales according to ReFED, and that percentage is the fastest lever any territorial program has.

Food waste is the public budget already paid that nobody claims

Fourteen percent of sales means a restaurant billing one hundred million a year throws fourteen into the bin, with product already purchased, transported, stored and often cooked. Total US food surplus added up to USD 380 billion that same year, of which 85% ended as waste, again per ReFED in its 2025 report. No development subsidy will hand an owner half that money, and yet recovering it demands no capital investment, only a scale, a log sheet and four weeks of discipline. Measure it in tons avoided per establishment per quarter; percentages invite creative interpretation, weight does not. Take the three figures in this block and you will see they point to the same redesign: 6.5% of regional employment, 70% informality and 14% of sales lost to food surplus according to the ILO and ReFED. A program that trains first and tries to measure later buys all three blind.

Employment, formality and waste together: what decision these figures trigger

The correct order inverts the sequence, and at Masterestaurant we instrument it this way: scale and dashboard first for four weeks, workshop afterwards, because a cook who has already weighed his own waste arrives at the training with a number of his own on the table. The difference is financial rather than pedagogical. Avoided waste funds the payroll burden of the first formal contract, and that contract is what sustains the twenty-four-month series you will later audit against real records. Concrete decision: shift 30% of the workshop budget toward prior instrumentation and require a baseline before the first disbursement. Every percentage point of food cost purchased inside the department multiplies the program's territorial effect, and with a sectoral food cost the National Restaurant Association places between 28% and 35% of sales, you are moving close to a third of the entire local gastronomic ecosystem's revenue.

Local sourcing: the figure that turns a restaurant into public policy

Run the numbers with two hundred establishments averaging two hundred million in annual sales: aggregate food cost lands near twelve billion a year, and lifting local sourcing from 20% to 35% redirects roughly one billion eight hundred million toward producers in the territory itself. No closing fair with press coverage moves that figure. What moves it is a volume-and-calendar agreement between fifteen kitchens and four producer associations, with specified grading and delivery frequency. Measure local sourcing as a percentage of food cost, backed by invoices, and publish it quarterly. The competency gap sits not in cooking technique but in reading contribution margin, which is why internal promotions stall at the walk-in door. A line cook reaches station chef through seniority, yet reaching executive chef requires defending a recipe cost sheet, tuning a menu to the optimal 28% to 35% food cost the National Restaurant Association sets, and holding a supplier conversation without conceding portion weight.

The skills gap blocking internal promotion, and its hidden cost

When that jump fails to happen, the establishment imports the role from another city, pays 40% more in salary and loses its line cook, who leaves for another kitchen chasing the promotion he never saw here. A territorial program teaching only recipes is financing the outflow. Teach costing using the establishment's real menu and measure how many internal promotions occur within twelve months. Suppose a program reports sixty-four jobs created in its final report, and eighteen months later social security records confirm only nine active contracts. Nobody lied, nobody simply looked again. Now carry the scenario to its consequence: if the program cost four hundred million, cost per declared job comes out slightly above six million and sounds reasonable in any committee, but cost per real job exceeds forty-four million and would not survive a single uncomfortable question.

What would happen if the municipality measured at twenty-four months instead of at the closing ceremony?

The paradox of this trade is that the program that looks cheap on paper turns out to be the portfolio's most expensive, while the one demanding prior instrumentation, slower and less popular during its first quarter, ends with the lowest cost per sustained job.

The clause resolving this fits in one contract line: payroll verification at twenty-four months, with payment conditioned on it. Three numbers decide whether your gastronomic local development program achieved anything, and none of them appears in the closing fair photographs. The first is 70% sectoral informality according to the ILO: your action is to demand an active social security payroll record at twenty-four months as the only valid evidence of job creation, never the workshop attendance list. The second is 14% of sales lost as surplus in foodservice according to ReFED 2024: your action is to hand out a scale and a dashboard before the first workshop and measure tons avoided per quarter, because that money is already paid and you only have to stop throwing it away.

The 3 figures you should tattoo on yourself

The third is the 28% to 35% food cost the National Restaurant Association sets: your action is to fix a local sourcing target as a percentage of that cost, verified with invoices, and publish it every three months. Start this week with the scale. The first one is sequence. A traditional program trains first and then tries to measure; the Twin Ecosystem instruments first and trains on data the restaurant is already watching on its own dashboard. When a cook weighs their waste for four weeks before the first workshop, training stops being abstract: it arrives with the cook's own number on the table. That inversion of order explains much of the gap in cost per job created. Second comes the unit of analysis. The classic approach counts beneficiaries; we count establishments with a live data series. Eighty beneficiaries scattered across forty kitchens produce forty incomplete data lines. Twenty-five properly instrumented establishments produce twenty-five complete series, and with twenty-five series you can already model employment elasticity against contribution margin in that specific territory.

Four differences that change the program's outcome

The third difference is the uncomfortable one: the traditional approach treats food cost as a matter of private efficiency, and it is not. Food cost above 32% in an independent restaurant compresses margin until formal payroll becomes unviable, and the owner solves it by going informal. The causal chain runs from a single recipe card to the municipality's informality rate, and that chain is what a serious LED program must intervene. Fourth is instrument permanence. An M&E system designed by the consultancy dies the day the final report ships. A platform the restaurant uses to place orders, cost recipes and run payroll stays alive because it solves the day. That is why 36-month evaluation exists in one model while in the other it is a phone survey with a 22% response rate.