Last week, Vox’s Matthew Yglesias reported that food is getting more affordable in America. His point is an important one: despite fear-mongering among conservatives, inflation continues to be quite modest, even if you look at a volatile sector like food.

The real economic challenges we face—weak labor force participation, stagnant wages, rising inequality, crumbling infrastructure—require accommodative monetary and (we wish) fiscal policy, not austerity.

But Yglesias’ analysis also misses something crucial: things are getting a lot harder for low-income Americans, starting at the dinner table.

The first problem is how we define affordability. Yglesias’ measure, food spending as a share of disposable personal income, is about budgeting. But what we really want to understand is the relationship between prices and income.

If I spend less of my paycheck on food, it doesn’t necessarily mean food got cheaper. In fact, between 2000 and 2013, food prices increased considerably more than the general price level: 41 percent versus 29 percent. What this means is that, relative to other things people spend money on, food is getting less, not more, affordable.

Smaller budget shares could also mean I’m spending more on something else. That something else is housing. In 1984 (the earliest year for which the Bureau of Labor Statistics has easily accessible data), the average household spent 15 percent of its budget on food. By 2013, the food share had dropped to 12.8 percent. For housing, budget share rose from 30.4 percent to 33.1 percent.

Poor households are the hardest hit by these trends. While food spending as a share of household expenditures dropped considerably for the top four income quintiles—from 14.9 percent in 1984 to 12.5 percent in 2013—for households in the bottom quintile, it didn’t budge, holding steady at 16 percent.

If the demands on the poor were proportional to those on the wealthy, low-income households would have an extra $450 a year.

If you look at which food prices have risen most during the past year, you’ll notice a perverse trend: the items low-income households spend most of their money on are those whose prices have risen the most. Meat—which, at 15.2 percent, accounts for the largest share of low-income households’ food budgets—has had the largest price increase of any food category: 7.7 percent since May 2013. Dairy products (4.2 percent) and fruits and vegetables (3.2 percent) have had the next largest year-over-year price increases. They rank fourth and second, respectively, in low-income households’ food budgets.

These categories are also the areas where the gulf between the poor and rich are the widest: compared to households in the top income quintile, households in the bottom quintile spend 1.2 percent more of their overall budgets on meat and 0.8 percent more on fruits and vegetables. If the demands on the poor were proportional to those on the wealthy, the typical low-income household would have an extra $450 a year. That’s nearly two weeks of work at the minimum wage.

The poor also fare worst in the housing market, spending nearly 40 percent of their budgets on shelter. (Bear in mind the federal definition of housing affordability is 30 percent of income.) Taken together, food and housing demands mean the households in the lowest income quintile have less than 44 cents of every dollar left to spend on all other needs.

That’s nearly a fifth less discretionary budget than all other households—and that’s in relative terms. In raw dollars, the average poor household spends $9,900 on non-shelter, non-food needs, compared with $32,200 among other households (including those in the second income quintile, who spend $16,300). It’s a vast divide.

To recap, there are two key points: (1) food actually isn’t getting cheaper; and (2) the poorest households disproportionately suffer from rising food prices.

Another way to think about the predicament is in terms of income. Since 2000, food prices have risen almost three times as fast as mean income for households in the bottom quintile. (It has also outpaced the income growth for households in every quintile, but less dramatically.)

Or we could consider food insecurity. If you want a bare measure of basic well-being, this is it. The incidence of food insecurity rose by a third during the Great Recession, and it hasn’t come down since. One in seven American households are food insecure, meaning they report three or more behaviors indicative of nutritional hardship, according the USDA’s Economic Research Service. Six percent are severely insecure (6+ behaviors). Fifteen percent of residents of the richest nation in the history of the world don’t have enough to eat–that is a stunning accomplishment.

To make matters worse (not that we really need to), none of this takes into account quality. It is well-known that cheaper diets tend to be more unhealthy. Though there is not much evidence to suggest the price disparity between unhealthy and healthy foods has changed much in recent decades, those with tighter budgets are more likely to end up with tighter belts as well (especially women and children). Access, convenience, and stress also play roles. In the long run however, the health costs of unhealthy eating far exceed the upfront price of better food, compounding disadvantage.

So the story is richer—or, perhaps more accurately, poorer—than Yglesias would have us believe. Nevertheless, the more textured analysis doesn’t contradict his argument; it reinforces it.

A smaller share of national income goes to food than ever before in our history, but underlying this macro trend are acute budget pressures on the poorest Americans. Putting the two together, we see that not only is inflation not a problem, we also have a powerful social welfare rationale for government stimulus.

One in seven American households are food insecure.

I’ll end on a positive note—one that demonstrates how the right kind of price increases can be a good thing. Economies are intertwined entities. Your spending is my income, and vice versa. If the recent run-up in food prices has a silver lining, it’s that it has been accompanied by higher wages in the food service industry, which is where many low-income workers earn their living.

Indeed, in 2013, food service accounted for two of every five low-wage jobs (defined as those where full-time, year-round employment yields less than the poverty line for a family of four). Between 2007 and 2013, the median hourly wage for food service workers increased 13.8 percent, from $8.02 to $9.14 an hour. Although the wage is still paltry, the rate of increase outpaced both that of overall median wages (11.7 percent) and inflation (12.4 percent).

That’s not to say higher wages for chefs and waitresses are primarily responsible for rising food prices. For the most part, food costs fluctuate with global commodity prices, which are subject to all sorts of volatility, weather and otherwise. But the wage-price nexus does point to a remedy: it only takes small increases in spending to generate large benefits in worker welfare.

How is that? Labor accounts for about a third of a typical restaurant’s costs. Consequently, even if restaurants passed the full cost of, say, a 20 percent raise on to customers, we would expect no more than 7 percent higher prices (one-third of 20 percent). In all likelihood though, the actual price increase would be far less, because higher wages can also boost worker retention and productivity, helping to offset costs elsewhere. Chances are, you wouldn’t even notice if a $5 burger started costing $5.25—but that’s enough to pay for a 15 percent raise for the guy flipping it.

The logic is essentially the reverse of David Card and Alan Krueger’s famous finding that increases in the minimum wage tend neither to decrease employment nor precipitate a price increase of the same magnitude. That is, price increases targeted at worker welfare result in more than proportional wage increases. In plain terms, what this means is that if the ongoing living wage campaigns in the fast food industry are successful, consumers will see little change, but workers will have significantly better lives.

Collectively, we’re spending less on food than ever before. For the sake of the poor, it’s time we started spending more.