Comment on the August 18 prediction: Are they reading my postings? Prices reset to $3.49 the next day, so the prediction was WRONG. However …

Sunday, August 25, 2024, 7 PM: The gas stations I follow the closest are on my commute to Allendale, and a week ago they were asking $3.46 a gallon. Then, on Monday, we got the “hike” to $3.49 (which I know affected other places a lot more), which did not seem warranted. By Wednesday, the stations were at $3.42, following one station that didn’t add the big three cents for the hike. Oil heading down last week also contributed, and today we are at $3.37, which is still too high if we do our usual calculations. So, I will predict, again, lower prices this week and no hike before the holiday weekend.

Meanwhile, our postings this summer about possible AI-fueled collusion and higher prices have turned out to be prescient, at least as far as real estate rental. Last week, the federal government filed an anti-trust suit against RealPage, the maker of the rent-setting algorithm we have been writing about. If there is something similar for gas prices, or aspirations to have such a software tool, this suit may make a difference. (Or, maybe this has all been my imagination.) -EA

Comment on the August 11 posting: I was indecisive about making a prediction, and thankfully there were no hikes in Michigan this past week that I am aware of.

Sunday, August 18, 2024, 5 PM: Let’s take a break from speculation about gas prices and artificial intelligence, and start over again with the basic calculation. We begin with an estimate of base wholesale price, which I have various sources for, and on Friday, that price was near $2.19. Meanwhile, ethanol is cheaper at $1.81, and regular blend is 90/10, which gives us about $2.15 a gallon. Now we add the federal tax (18.4 cents per gallon) and my educated guess of add-ins from middlemen and credit card fees (18 cents per gallon), bringing us to $2.61 or so. Now on to the Michigan state taxes: 6% sales tax and 30 cents per gallon, bringing us to about $3.07. That is what I call the 0-cent margin price, and because the base wholesale price is fuzzy, so is this number. But, it is a reasonable floor for gas prices in the state of Michigan.

OK, let’s take a look at actual prices in the Grand Rapids area, via Gas Buddy: In Sparta, often a place for the best prices, there is a station posting $2.97 this afternoon. Woo hoo! A bit of a gas war in Greenville, in the $3.10’s. Looking at higher prices, there are a lot of $3.40’s and $3.50’s in the area.

There is one other twist for the mid-August to mid-September time period. Historically, the retailers are in no rush to restore prices and margins at this time of year. I’m not sure why, but it may be related to switch over to cheaper winter blend gasoline in mid-September. Given that, despite some stations down near $3, I am predict we’ll continue to see prices dropping the next several days, and crossing my fingers that we’ll make it to next weekend without a hike. -EA

Comment on the July 27 prediction: Regarding sliding prices, for the most part CORRECT, although there have been scattered hikes around the Midwest.

Sunday, August 11, 2024, 7 AM: It looks like wholesale prices have dropped 50 cents a gallon the past two weeks, as the refinery in Joliet has come back online. Average retail prices have been working their way slowly lower, and I would like to see that accelerate this week. We are down to $3.29 in Greenville, MI (which is near Grand Rapids) and I’d like to see many more stations adopt that price.

But, I can’t make that prediction today, because of the higher-margin world we have been living in this year. That brings me to my last post on July 27, I wondered whether some sort of artificial intelligence program is now being used to set gas prices — collusion via algorithm. I had no direct evidence for this idea and I was just trying to suggest a reason why playing the Gas Game has seemed different this year. Not surprisingly, I am not the first person to think about this question when it comes to gas prices, and in fact, a research article came my way this weekend that looks at algorithmic pricing (AP) of gasoline in Germany, starting in 2017. Here is an excerpt that I think gives away the story:

“Simulation results … suggest that it can take a long time for algorithms to train and converge to stable strategies. During this time, algorithms may learn to punish competitors for reducing prices or other tacitly-collusive strategies.

We find evidence consistent with this. Margins do not start to increase until about a year after market-wide adoption, suggesting that algorithms in this market learn tacitly-collusive strategies. We also examine the pricing behaviour that emerges in markets where both duopolists are algorithmic adopters. We show that in a market where both duopolists adopt, a station is more likely to respond to a rival’s price decrease with an immediate price decrease of their own. There is no comparable change in the propensity to respond to price increases by a rival.

We also find that when both stations adopt algorithmic pricing, the duopolist setting higher prices is less likely to undercut the duopolist setting lower prices. The timing of these effects is consistent with the timing of the price and margin increases. Altogether, these findings provide further evidence that adoption affects competition and they suggest that the algorithms learn that undercutting will not be profitable, since the undercutter will always be followed to the lower price by its rival.” [bold is mine]

Duopoly … two major players in the market. Hmmm, sounds familiar. If we are dealing with an AP-powered duopoly in the Midwest, and these results from Germany are transferable, that would suggest from my data that there was some training of algorithms in 2023, leading to larger margins in 2024.

Again, this is all speculative, but I would not be suprised if some economist is working on reproducing the German study with the markets I follow.

Readers may also be interested in this new article in The Atlantic on the subject of AP. -EA

Comment on the July 18 posting: We got through last weekend without a hike, and I thought that meant we were safe. But, prices rose last Tuesday to $3.99, which meant “If so, I think they’ll keep it under $4” was accurate.

Saturday, July 27, 2024, 3 PM: As noted here and elsewhere, a violent storm nearly two weeks ago took out the second-largest refinery in the Midwest. As far as I can tell, that refinery is still offline, but a combination of imports from other parts of the country and a selloff in oil prices last week has started to put downward pressure on wholesale prices. Consequently, depending on where you drive, you can find gasoline at $3.89 or even lower. I predict we’ll see prices continue to slide, and once Joliet is back up and running, maybe $3.25 somewhere?

While recent price action does seem to match our idea of supply and demand, I’ve commented earlier this year that price dynamics have not been following our Gas Game script for a while. Prices aren’t dropping as low before a hike as they would in the past, and the hikes are higher above wholesale than they used to be. It has made me wonder if there is some other cost being added in that I am not aware of. While sitting on the beach recently, I came up with a different theory, tied to greed: AI.

In a 2022 article about the pricing of apartment rentals, ProPublica described how landlords in larger cities have been using an AI-infused program called AI Revenue Management (a.k.a., YieldStar) to make decisions about the rents they charge. Key excerpt: “The company had been seeking occupancy levels of 97% or 98% in markets where it was a leader. But when it began using YieldStar, managers saw that raising rents and leaving some apartments vacant made more money.” YieldStar works by collecting rental data from throughout the market and using AI to make recommendations.

I have no evidence for this, but what if one or more of the retailers have started using some sort of AI Gasoline Price Management software this year? This software would constantly monitor prices throughout the Midwest (like GasBuddy does via crowdsourcing) and then make some calculation that higher prices might be worth the trade-off in number of customers or positive reputation. Particularly when you consider the way competition works, or doesn’t work, in retail gasoline, this might be a good money-making strategy.

If that is the case, then note that the people who own or run the retailers aren’t talking to each other — in other words, there is no obvious collusion. But, to quote the ProPublica article, “Is it OK for a guy named Bob to collect confidential price strategy information from all the participants in a market and then tell everybody how they should price? If it isn’t OK for a guy named Bob to do it, then it probably isn’t OK for an algorithm to do it either.” In other words, the collusion would occur in the software.

Like I said, I have no evidence for this, but starting to wonder why this year of ChatGPT has also been a year of a change in gas price dynamics.

On the other hand, if I just suggested an idea for a new company to sell to the retailers, then I’m sorry. -EA

Thursday, July 18, 2024, 8 PM: If you haven’t heard, that nasty storm earlier this week knocked out the refinery in Joliet, Illinois. While the plan is to get the complex running again this weekend, this loss of supply has led to some short-term jumps in wholesale markets in the Midwest. Retail is following, and prices jumped to $3.79 in Michigan today. Could we see another hike before Sunday? Yes, we could. If so, I think they’ll keep it under $4. Will we see that hike? No prediction. -EA

Saturday, July 13, 2024, Noon: Greetings from Chicago, where gas prices are typically a half-dollar or more higher than the rest of the Midwest. This morning, I ran in the Big Ten 10K, and now it is time to discuss the latest with gas prices.

After a quiet first week of July (hooray!), we saw prices reset last week, up to around $3.75 a gallon. It was another one of those inflated margin hikes, this time led by Big Red, as I watched Big M on Lake Michigan Drive drag their feet for at least a day before matching the hike. On the other hand, the Costco that I monitor has gotten into the Game, posting a price of $3.43 yesterday, versus $3.06 a month ago, even though wholesale prices are practically the same today as they were then. Meanwhile, diesel prices are about a half-dollar higher than a month ago.

So, as I’ve written previously this summer, something else is in the mix right now. Even though wholesale prices have been stable, that may reflect that both supply and demand are changing in the same way. But more demand makes it easier for the stations to keep prices higher.

The plan this week? Stay away from stations asking more than $3.59 a gallon. Neighborhoods that include a K&G are the best. -EA