Are You Overpaying for Taco Bell? There's a Good Chance
The Supreme Taco Party Pack Index of Price Gouging
Key Takeaways
I looked at prices of the Supreme Taco Party Pack (12 tacos, about $25) across ~7,200 Taco Bell locations in 45 states.
There’s not a huge spread — most locations are within 15% of the average price nationally (though in a fast food context, 15% might be significant).
Prices are moderately correlated to cost of living in an area.
Looking at 300 cities across the country, almost exactly half overpay relative to the city’s cost of living — but usually by no more than 10%.
Taco Bell in Hawaii, California, and Massachusetts is 20-60% cheaper than it should be.
Some towns in upstate New York and Oklahoma are getting gouged, paying 20-30% more than they should be.
With more time, I would a) look across more menu items b) compare 2023 vs. 2024 prices/distributions, and c) set up a website where users could see how the price in their city compares to other cities.
Interpretation of the chart: Relative to the cost of living in thee state, the average Oklahoman pays 20% more for their Taco Bell than they should. In Hawaii and Massachusets, Taco Bell is a real value play.
Intro
Recently, in Charlotte, North Carolina, a friend and I went to Taco Bell, as suburban youths do. I looked at the prices and said to my friend “I bet this would be much more expensive in New York!” He replied, “really, how much?” I realized that I didn’t really know. A dollar, maybe two? Three?
So I decided to find out. Big chain fast food is hyper-standardized. A Supreme Taco Party Pack (this is a real thing) should be the exact at each Taco Bell location, with the only major differences attributable to geography. If this sounds familiar, it’s the same logic behind the Big Mac Index.
So, I set out to see
a) How much spread there is in prices.
b) How much does cost of living explain that spread — expensive places should have more expensive Taco Bell
c) Are there cities in which Taco Bell prices are more expensive than the cost of living index implies? God forbid, are you overpaying for your Taco Bell?
While I scraped data across all menu items, the benchmark for this analysis is the humble Supreme Taco Party Pack: a ~$25 menu item; it’s a pack of 12 tacos for you and your friends to enjoy together while reading this.
Process
Most readers probably aren’t super interested in the nitty-gritty of scraping, so here’s the high level: I found a company-provided listing of Taco Bells online, followed that to the menu pages for each location, and pulled the data for all of them.
The extremely helpful https://locations.tacobell.com/:
A menu page for a location. Yes, North Pole, Alaska is a real place.
Once the code was together, a full scrape of all locations and their menus took about 45 minutes. Once I had all the data, I did a quick survey of missingness. A handful of states couldn’t be scraped adequately (or at all), so I omitted those from the analysis.
In the end, we have the prices across almost 7,200 locations spread across 45 states.
Results
Spread in Prices
Here, we histogram the 7,200 prices of the Supreme Taco Party Pack:
There are two things that I take away from this chart:
The spread of prices is fairly tight. 84% of the prices fall within 15% of the mean. So, if you order a taco party pack at a random Taco Bell somewhere in the country, there’s a pretty good chance you’re not paying dramatically more than the national average
Individual locations adhere closely to pricing guidelines. More than 1,000 locations price their Party Pack between $29.50 and $30.49. Nearly another 1,000 are between $25.50 and $26.49.
Overall, this makes sense, and fits with what we might think about fast food — should be fairly similar everywhere.
Correlation to Cost of Living Indices
That said, there is going to be some difference right? And if everything is standardized, we might think deviations in price should be largely attributable to cost of living in a place — let’s take a look then at how well cost of living indices explain prices:
State Level
I pulled a cost of living index provided here. Then, I aggregated mean prices for the Party Pack and regressed those against cost of living
There’s a moderate correlation! As states get more expensive, they generally have more expensive Taco Bell. The regression line indicates that, on average, for a 1 point increase in a state’s cost of living, prices go up by $0.07.
What’s interesting is to look at are the states in which the cost of living is quite different, but average prices are similar. People in Massachusetts pay the same for their Taco Bell as people in Missouri, yet Massachusetts is 68% more expensive! While Hawaii is notoriously expensive to live in (as detailed recently in a great John Oliver segment), apparently Taco Bell takes pity on them and charges them the same as their fellow Alaskans.
On the flip side there are many states with similar cost of living that have very different prices — Oregonians pay about $5 less than their counterparts in New Hampshire despite similar costs of living.
Overall, it looks like Taco Bell is using a pricing strategy that generally tracks cost of living but is also driven by many other factors, as there is a lot of variation.
Of course, the cost of living in a state can vary dramatically based on whether you’re in a city or a rural area. So it makes sense to look at the city level as well.
City Level
I pulled a cost of living index for 500 cities from here. 300 of those cities matched the Taco Bell cities exactly. Then, I took the average price for each city and plotted that against cost of living. The labels become pretty unreadable with that many points, so I just labeled the handful of top cities.
Here, the correlation is a little weaker at 0.43.
Perhaps even more prominently than the state plot, we see how a lot of the cities have either the same price but very different costs of living, or vice versa. And those poor college kids in Auburn and Ithaca are actually paying more than people living 2,400 miles away in Honolulu.
The fact that we see many of the dots overlap at multiple levels on the graph (i.e. the exact same price) further reinforces the fact that locations adhere quite strictly to pricing guidelines.
It feels like those handful of cities with a cost of living more than 125 could be acting as leverage points skewing the regression line, so let’s try removing those and see what happens:
It doesn’t change much.
Given that correlations are meaningfully weaker at the city level, my overall takeaway is that cost of living only explains some of Taco Bell’s pricing. There are likely many hyperlocal factors that play a role (e.g. taxes, local and state wage ordinances, supply chain stuff)
Are you overpaying?
That’s all fine and good, but tell me what I care about — am I overpaying???
At a simple level, we might say that if you live in a city that scores a 110 on the cost of living index, your Taco Bell order should be 10% more expensive than that of a city that scores 100 on the cost of living index.
I look at percentage deviations from the mean for each. E.g, if your city is 15% more expensive than the national average, but the cost of your party pack is 20% more than the national average cost of party packs, you’re paying 5% more than you should! Highway robbery!
State Level
A handful of states are paying meaningfully less than they “should” be, but most are getting slightly overcharged!
Hawaii, Massachusetts, and California are making out like bandits, with their Party Packs being more than 20% cheaper than cost of living indices would imply. And Oklahomans are getting gouged! 20% mark ups!
The distribution is a little more easily visualized with a histogram and the numerical distribution:
75% of states pay <= 7% premium for their Taco Bell. Half of them pay basically no mark up or get it for cheap. And then if we look at the middle 25th to 75th percentile, we see it’s basically a wash, and Taco Bell is priced about where it should be, to a touch higher. That’s not too bad! It’s at the extremes that things get wild.
City Level
A different picture! It seems like almost an even split between cities overpaying and underpaying. Let’s look at a histogram:
We still have the same conclusion that 75% of cities pay <= 9% of a premium and that the middle between 25th percentile and 75th percentile pay a little less to a little more than they should be.
And finally, the most overpriced and most underpriced cities for Party Packs:
If you ever find yourself in Ogdensburg, NY, don’t go to Taco Bell. If you’re in Napa and have the munchies after a wine tour, Taco Bell is a solid option to feed the gang.
Limitations
Looking at only one menu item: I chose one $25 menu item to zoom in on. Whether a single taco is $1.50 or $1.99 is a little less interesting, at least to see visually, and kind of hard to think about. If my taco is $1.50 and yours is $2.30, that’s a 50% difference but….is that meaningful? In a fast food context, perhaps. I’ll leave that question for my hedge fund friends who cover quick service restaurants.
Unclear what subset of totals stores we have: I believe that there are locations not listed on https://locations.tacobell.com/. Multiple articles like this one and this one mention there are about 15,000 Taco Bells in the US. Yet if we sum up the numbers listed on the state page, we get a touch under 8,000 locations, and the Yum brands page lists 8,000 locations:
It’s unclear what’s going on there (maybe we only have company-owned stores). But seems like we have a substantial portion of locations in our data set.
Chosen Cost of Living Index: Building a Cost of Living Index from scratch was a little much for me. So I have to hope the ones that I pulled from a Google search are reasonably well-constructed, but I’m at the mercy of their methodologies. Additionally, the cost of living index for cities and for states one come from two different companies, so that could lead additional disparities.
Many more factors in pricing: The fact that cost of living is only moderately correlated checks out. Plenty of hyperlocal factors like supply chains considerations, local wage ordinances, and variations in rents are probably inputs to price that we can’t capture.
Future
Compare YoY prices: I did this scrape once in 2023, completely forgot about this project, and then re-ran the scrape in 2024. It’d be interesting to see how prices and distributions have changed over that time.
Consider every menu item: With some creative use of groupbys and loops, I could do these analyses across every menu item instead of just one. It’d also be interesting to consider the menu items in aggregate (e.g. the average/median price at a location) and to use that for the basis of our analysis.
More nuanced geographic consideration: It’d be interesting to compare various MSAs to each other. Rural areas and suburbs vs. large cities, zoom in on MSAs, on metropolitan areas vs. micropolitan areas (more data cleaning)
Create a web app displaying the data: With something like Plotly Dash, it could be fun to have a website where people can go, search their city, and see how their menu item prices compare to those of other places.
Technical Details
github: https://github.com/arjungup740/taco-bell
Scraping Data
Conveniently, Taco Bell has a nice simple website that you’ve probably never used.
Each location has a store ID, that’s passed to a pretty straightforward API. Now, we just need store IDs en masse.
It turns out that, within each state page, there are more links listing each city in that state and their Taco Bells.
Even more helpfully, embedded within the HTML of each location page, is a store ID.
So we just had to loop through all the individual pages to grab the store IDs, and then pass each of those to the API. Figuring this out and doing it took a bit of effort, but it was very doable.
Interestingly, most of the store IDs for Vermont and Maine were of a different format (often had letters in the store ID while the rest were purely numeric) and those locations didn’t have corresponding menu pages. So it seems like some locations have closed but the pages haven’t been updated.
If you made it this far, thanks for reading!
It's a very good point. For the sake of digestibility I zoomed in on that one menu item. In future, could look at these charts across all items, or the average of a basket of a few core items for each store
Very interesting. I'm curious if the Party Packs themselves might be getting priced differently compared to more commonly ordered, non-Party items on the menu, such as a basic soft taco or a long-standing menu item.