The price of guaranteed execution: automated market making versus bilateral peer-to-peer matching in a network-constrained local electricity market
This paper demonstrates through co-simulation on a real-world low-voltage feeder that automated market makers (AMMs) significantly outperform bilateral peer-to-peer (P2P) matching in local electricity markets by delivering up to 27% higher prosumer welfare and drastically reducing settlement complexity, particularly when network constraints limit the availability of counterparties.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the quiet corners of our electrical grids, a new kind of neighborhood economy is trying to take root. Instead of every home simply buying power from a distant utility company and selling excess back to them, neighbors are beginning to trade electricity directly with one another. This concept, known as peer-to-peer trading, promises a more efficient and local energy system. However, it faces a stubborn, practical problem: timing. On a sunny afternoon, a neighborhood full of solar panels might all produce extra power at the exact same moment. If everyone is selling at once, there is no one left to buy. The market stalls because the necessary partners simply do not exist. To solve this, some researchers have proposed borrowing a tool from the world of digital finance called an automated market maker. Think of this not as a person, but as a smart, always-open vending machine that stands ready to buy or sell energy whenever a neighbor needs it, ensuring a trade can happen even when no other neighbor is available. The big question has always been whether this machine is truly better than waiting for a human match, or if it simply adds unnecessary cost.
A team of researchers from Colombia set out to answer this by building a detailed digital twin of a real neighborhood. They did not just guess how these systems would behave; they simulated a full year of life on a low-voltage electrical feeder, a specific type of local power line that serves 118 homes equipped with solar panels and batteries. They created four different scenarios to test. In the first, neighbors traded nothing and simply bought or sold from the main grid. In the second, neighbors tried to trade directly with each other, but if they couldn't find a partner, the trade failed and their batteries sat idle. In the third, they tried the same direct trading, but this time they allowed the neighborhood batteries to actively participate in the search for a partner. Finally, they tested the automated market maker, a system backed by those same batteries, which stood ready to trade instantly without needing to find a specific partner.
The results revealed a clear winner, but also a surprising cost. The automated market maker delivered significantly more value to the neighborhood than the direct trading methods. Specifically, it generated twenty percent more annual benefit for the homeowners than the best version of direct trading, where the batteries were actively trying to find partners. When compared to a version where the batteries did nothing, the advantage grew to twenty-seven percent. This benefit was not a one-time fluke; it grew steadily as the neighborhood became more crowded with solar panels and the likelihood of finding a trading partner dropped. In the worst-case scenarios, where no neighbors were available to trade at all, the automated system provided nearly double the benefit of the direct trading approach. It worked because it did not wait for a match; it simply executed the trade.
However, this guaranteed execution came with a price, and the researchers measured exactly what that price was. The automated system did not always serve every single request. Because the battery backing the system has a limited capacity, it could only fully satisfy the neighborhood's needs in about forty-three percent of the active trading intervals. In the other moments, it had to ration the energy, serving only a portion of what was requested. This rationing happened most often precisely when it was needed most: during times when no other neighbors were available to trade. Additionally, the automated system caused a small shift in the price of electricity, an average impact of 2.6 percent, though it never pushed prices outside the safe limits set by the local utility tariffs.
Perhaps the most unexpected discovery was not about the price or the energy, but about the paperwork. The researchers found that the two systems differed wildly in how many individual transactions they required to clear the market. The direct trading method, where neighbors pair up one by one, required roughly seventy-seven separate settlements for every active trading interval. The automated market maker, by contrast, handled the same volume with just two settlements. This created a massive difference in efficiency: the automated system could absorb settlement costs forty-nine times higher than the direct trading method before those costs ate up all the profit. In a world where every transaction might carry a fee, this difference suggests that direct trading might be too expensive to scale, while the automated approach remains viable.
The study also corrected a common misunderstanding about how these markets work. The researchers initially suspected that the automated system's success was simply because it had access to batteries while the direct trading system did not. To test this, they added the batteries to the direct trading system and let them bid for trades just like the neighbors. Even with the same physical resources, the automated system still won by twenty percent. This proved that the advantage came from the method of matching itself, not just the presence of extra equipment. Furthermore, the researchers found that making the neighborhood larger did not help the direct trading method. Even as more people joined and the chances of finding a partner theoretically improved, the automated system's lead only grew wider.
Ultimately, the research shows that for local electricity markets to function reliably, especially when everyone's solar panels are producing at the same time, a system that guarantees a trade is superior to one that relies on finding a partner. The automated market maker achieves this by taking on the risk of holding inventory, a role that requires a small fee to keep the system running fairly. While the system does not solve every problem and must sometimes ration energy, it provides a robust, efficient, and scalable way for neighbors to trade power without getting stuck in a deadlock. The study concludes that the cost of guaranteed execution is real, but it is a price worth paying to keep the local grid moving when the sun is shining and the neighbors are all selling at once.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.