What will happen to Mamdani's grocery stores?

We finally have some more details on what NYC’s Mayor Zohran Mamdani has in mind for his government grocery. He has announced his plans for city-owned, contractor-operated grocery stores in New York.

Today, I am proud to announce a collection of essential staples that will be predictably 30% cheaper at all five of our city-run grocery stores. This core set of goods will include all fresh produce, meat, and seafood, along with 20 other essential items like cheese, milk, and bread.

Here’s how it will work. Once a month, our five city-run grocery stores will set prices for this core set of goods at 30% below typical retail prices. No exceptions, no gimmicks. The savings will last for the entire month.

I could spend this post cataloging all the ways a municipal grocery store is a dumb idea. That’s boring.

So let’s take New York’s plan seriously and think through what we always think through. Who pays for the discount? What are the equilibrium responses? Where does adjustment occur when it cannot occur in the prices of those 20 items? Let’s think through the usual stuff: competition, transactions costs, fixed vs. marginal decisions.

Let’s start with a few of the details. The 30% covers a subset of items, not every item. So it’s not right to compare the 30% to the overall grocery margins, which are in the single digits. Also, this is not a price control. Nothing binds private stores; they can set their own prices. The price-control analogy will become useful later, but a low checkout price by itself does not tell us whether the store receives enough in total to keep the shelf stocked.

The city will provide the locations. That means low- or no-cost real estate, initial building support, and relief from rent and property taxes. So there will be a huge subsidy on that portion, and bids to run the stores will be judged on what subsidies are requested, but we don’t know the formula. This matters because many of these subsidies are about the fixed costs of the store, while pricing (and the rules around the 30%) are related to marginal costs. So we will need to be careful here. Lots of moving parts.

Competition for the bundle

Let’s start by assuming that the supported operator, which I’ll just call the government store, and the ordinary private grocer use the same resources to build their basket that consumers care about. I’m not going to assume that the government store is inherently inefficient at the start, but we’ll come back to that later.

It’s maybe an okay starting point. We don’t need to assume the city is learning how to price groceries. It’s going to hire private contractors to do that and give them some residual-claimant rights, so they have an incentive to learn. They’ll also likely have a scale advantage that other stores do not, so let’s start by assuming they are equally efficient.

As always, start with the simple model. Imagine two stores. One private store becomes the government store, and the other remains private. Assume every household consumes a fixed grocery basket, so we’re not going to think about substitution across goods. Consumers are going to go to whoever supplies the cheaper basket relative to their transportation cost.

We have core items whose prices are “forced” low and the rest of the items. Suppose half of the $100 on a normal trip is on those core items, and the city is going to tell the store to cut that by 30%. The dollar value of that portion falls to $35. If nothing else changes, the $100 trip now costs $85. Yay, everything is cheaper! The advertised 30% discount has produced a 15% discount overall.

But the operator doesn’t sell a single product. It sells lots of products together. Half are regulated, but the other half are flexible. This operator is still competing with the private store. What is it going to do? It’s going to raise prices on the non-core half until the total basket is back to $100.

In that world, nothing has changed. Every basket is the same price. Shopping behavior is the same. Even if some staples look like they’ve changed, consumers are no better off in this world because what they actually care about is the bundle.

Also notice that nothing on the subsidy side is really coming in here in terms of pricing. I’m not assuming that the government store made a loss and therefore is getting a subsidy to cover that loss. It’s just a direct transfer in the form of cheaper rent from the taxpayer to the store operator.

But none of that gets passed through to consumers.

Competitors respond

Now let’s relax the fixed-basket assumption. Suppose consumers are a little more responsive and non-core pricing isn’t going to undo the full discount, so the supported store’s total basket really does get cheaper. We can think of a variety of ways in which that happens. One possibility is that people shift toward the discounted core products, so the average bundle contains more goods with lower margins. The government store then has lower prices on average.

Again, let’s think of a simple model. Imagine it’s a spatial competition model. One store sits at each end of the neighborhood. Consumers differ in which store is more convenient. The supported store has cut its basket by 10%. The rival is going to respond. It can either keep its price unchanged and lose customers, or cut its own price and keep more of them. Suppose the government store cut its basket by $10, and then the other store is going to respond by $5.

Sounds great! Everyone gets cheaper groceries! People going to the government store get the larger reduction, while people who stay at the private store still benefit from the competition that has driven down its prices.

But we already know that these private stores have thin margins. With recurring fixed costs, it’s possible some are eventually driven out of business. In that case, this competitive pressure is weakened because the government store no longer has to compete with the private store. It can’t exercise its market power by raising prices on the core items, but it could still exercise its market power on all of those other goods.

The subsidy is real

So far, we basically ignored the subsidy. Think of it as just a transfer. Let’s dig into that a little bit more.

We have a supported operator that we’ve assumed has the same costs as the other firms involved. In the market, it’s just been told to charge less, and maybe it can’t make it all up on those other goods. Fixed contributions toward the site buildout and tax subsidies can help keep it operating, but they do not necessarily change the incentive to sell another basket.

This really turns the pricing into something like a two-part tariff, where you have kind of a fixed cost part of the pricing, and then you have your marginal cost pricing on the goods. Think of Costco as one example. Costco collects a membership fee and then charges very low prices inside the store. Maybe this is essentially what New York will do, but instead of membership fees, that fixed cost is coming from the government. After all, there’s a five-year plan that provides $70 million to go towards these sites.

We can see fixed vs. marginal in the profit function:

π = (P + s − c)q + G − F.

The store charges P, sells q baskets at marginal cost c, receives a fixed payment G and a per-basket payment s, and they themselves pay fixed cost F. They key is that the fixed payment doesn’t enter the pricing decision. It’s a sunk cost. Sure, it determines whether the operator opens (I’m fine assume it will be sufficient for the stores to open for a little bit), but it does not make the next basket, the next q, more profitable.

The way to change that is through s. So far, we don’t know what that subsidy looks like. But then that changes behavior as well. That changes the full equilibrium. In a model like Hotelling, another dollar of s lowers the government store’s price by say 67 cents and its rival’s by about 33 cents. That exact split is an artifact of symmetric Hotelling but the general point is that the fixed cost type support changes participation, or the extensive margin. Marginal support changes the incentive to expand/maintain quantity, which is the intensive margin. Always gotta think about extensive/intensive.

If that’s all we have, then again, we’re in the world where it’s just a transfer from the taxpayer to the store operators. Competition may make it optimal for the government store to pass some of the subsidy through to consumers rather than keep all of it.

Overall, we have a combination of fixed and per-basket or per-unit payments from taxpayers. That produces three distinct outcomes that often get collapsed in political discussions of subsidies:

• The amount transferred to the operator

• The amount that store shoppers receive—the pass-through rate

• The price cut at competing stores

The details will determine who captures each of these.

Will the stores be as efficient?

The elephant in the room is: why would these supported operators be more or less efficient? They could be more efficient because they’ve been given a scale or a chance to reach a scale that’s generally not permitted in New York City. After all, New York City has kept out stores like Walmart, and these will presumably be five bigger stores. We also have all of the usual reasons for thinking that the stores would be less efficient.

To what extent is there an inherent contradiction here? If you provide only a fixed-cost subsidy and let the operator maximize profits, then it has an incentive to be efficient. In that world, we go back to the case above, where most of the fixed subsidy is captured by the operator. Competition may pass some of it through to shoppers, but the operator will try to retain the savings.

Otherwise, if the operator can’t maximize profit, or if the government claws back profits that it considers too high, then the incentive to minimize costs is weaker. You’ve created a tension: letting the operator claim the gains from lower costs encourages efficiency, but it also lets the operator retain some of those gains rather than pass them through to consumers.

This will really depend on the details, so we will see.

What about shortages?

Notice I haven’t said anything about shortages or waiting in line, the usual things that people think about when they think about government-run stores. Let’s imagine that consumers go to the government store to get milk. Maybe they make one big trip there a week to stock up on the lower-priced, core items.

What incentive does the government store have to keep the shelves full? After all, the operator likely will lose money on the core items, and needs to make it back on the rest of the basket. But maybe it needs to have the core items to attract the shoppers for the rest of the basket. No one will show up to a grocery store with no milk and bread. So there is some incentive to keep shelves stocked.

Again, think about margins of adjustment. You know we love to bring up Barzel, but his ideas are key here. Measurement costs are real. He starts from the idea that what is sold is not one thing with one price. A carton of milk is a bundle. Yes, we all think about the (posted) price. That’s what all the emphasis is on. But we’re talking produce here. It doesn’t take a lot of thinking to recognize that quality matters too. But it’s also things like selection, checkout time, and even the probability it will still be on the shelf when you arrive. It’s just not possible to force the operator to do all of these correctly. That’s an aspect of all contracts, not just government ones. Writing and enforcing a contract over every attribute is costly, so contracts tend to be precise about the attributes that are easy to measure and looser about the rest.

The contract can be precise about the posted price. That’s what Mamdani is promising on. What about the other margins?

Suppose the contract requires the operator to “fix” the posted price. That’s the price control analogy, but it’s quite loose. That’s forced low, so another sale does not cover its cost. The operator then has an incentive to adjust on those less-measured attributes.

Let’s not assume shortages. It could be that other margins adjust. Less reliable availability or worse service may reduce demand enough, so we don’t actually have shortages. YAY! NO SHORTAGES! (Ignore for the moment that there are no shortages only because quality has fallen enough to offset the lower price.)

But if there are still more shoppers than groceries, Barzel’s earlier work on rationing by waiting tells us what comes next. Shoppers will then have to bid through another mechanism, in this case, bid with time. The line grows until the full price (price + waiting for simplicity), on the margin, is high enough to ration the available groceries. The market clears one way or another.

This is the standard deadweight loss of rationing. The time cost is wasted rather than being transferred to the store or anyone else.

So that brings us back to the opening. I still think this will end up displaying all the normal reasons to think a municipal grocery store is a dumb idea: bad pricing incentives, inefficient transfers, and all that stuff. At the end of the day, we can’t eliminate trade-offs and scarcity. This is a competitive industry. The discount is unlikely to come purely out of profits.

So the cost of the discount will show up in other grocery prices, taxpayer subsidies, product quality, or time. There’s no way around that.