Industrial Park Land-Use Composition and Occupancy Performance: Explaining Nonlinear Determinants of Industrial Land Absorption in South Korea

Imagine you are a city planner, but instead of drawing blueprints for houses, you are designing giant, empty neighborhoods specifically for factories. In the world of urban planning and economics, there's a long-standing debate about how to make these "industrial parks" successful. For decades, the go-to strategy was simple: build more land, make it bigger, and hope companies move in. It's like baking a massive cake and assuming that if you just make it huge enough, everyone will want a slice. But what if the cake is too big, or the wrong flavor, and it just sits there getting stale? This is the problem of "land absorption"—the fancy term for whether the businesses actually show up and fill the space. Researchers have long wondered if it's the location of the park (like being near a big city) or the inside design of the park (how much space is for factories vs. roads) that really matters.

Now, enter a team of researchers from Yonsei University in South Korea who decided to stop guessing and start measuring. They looked at over 1,100 manufacturing parks to see what makes them full or empty. They used a super-smart computer brain (a type of artificial intelligence) to find patterns that human eyes might miss. They discovered that it's not just about where the park is, but exactly how the land is sliced up inside. They found that there are strict "sweet spots" for how much space should be dedicated to actual factories versus roads and parks. If you go too far in either direction, the park stops working. It's a bit like a video game where you have to balance your character's stats; if you put too many points into "defense" and none into "attack," you lose the game, no matter how cool your armor looks.

The Great Factory Park Mystery

So, what did these researchers actually do? They took a massive dataset of 1,161 industrial parks across South Korea and asked a simple question: "Which ones are completely full of companies, and which ones are sitting half-empty?" Instead of using old-school math that assumes everything happens in a straight line (like "more roads = better"), they used a high-tech tool called XGBoost, combined with a method called SHAP. Think of XGBoost as a detective that can spot complex, hidden clues in a messy room, and SHAP as the magnifying glass that explains why the detective found those clues.

Their main finding is a bit of a plot twist. For a long time, people thought that if you built a park near a big city with nice schools and hospitals, companies would flock there. But the data suggests that's not the whole story. The most important factor is actually the internal recipe of the park itself. Specifically, the mix of land types inside the fence matters way more than the fancy amenities outside.

Here is where the "sweet spots" come in, and this is the most critical part of the story. The researchers found that the relationship between land use and success isn't a straight line; it's a bumpy rollercoaster with specific thresholds.

First, let's talk about the Factory Land Ratio. This is the percentage of the park dedicated to actual manufacturing buildings. The study found that if this ratio is below 0.65 to 0.70 (meaning less than 65-70% of the land is for factories), the park struggles to get companies to move in. It's like trying to host a dinner party in a house where the kitchen is too small; people just don't want to come. But once you cross that 0.65–0.70 line, the park suddenly becomes attractive. However, there's a catch: if you keep adding more factory land past a certain point, the benefit starts to flatten out. It's not "the more the better"; it's "just enough to get the job done."

Second, and perhaps even more surprising, is the Public Facility Ratio. This includes roads, green spaces, and other non-factory areas. The researchers discovered a sharp "cliff" here. As long as public facilities make up less than about 0.28 to 0.30 (28-30%) of the land, they help the park by making it accessible and nice. But the moment you cross that 0.28–0.30 threshold, the park's performance crashes. Why? Because you've squeezed out too much space for the actual factories. It's like building a beautiful, wide driveway for your house but forgetting to build the garage; the house looks great, but you have nowhere to park your car. The study explicitly rules out the idea that "more public space is always better." In fact, too much of it actively hurts the park's ability to get filled up.

The study also looked at other factors, like how old the park is (maturity) and how many other parks are nearby (agglomeration). They found that older parks tend to do better, likely because they have a reputation, and having other factories nearby helps. But when they looked at the "host city" factors—like how many libraries, museums, or high-rise apartments were nearby—they found these had very little impact on whether a specific park got filled. This suggests that a park's success is determined more by its own internal design than by the general vibe of the city it sits in.

Why This Matters

The authors argue that for a long time, governments have been making a mistake. They've been focusing on "quantity"—building more and more land, hoping that supply creates its own demand. But this study suggests that approach is like throwing spaghetti at a wall and hoping it sticks; it's wasteful and often leads to empty, ghost-town parks.

Instead, the paper suggests a shift toward "quality" and "balance." If a park is struggling, the solution isn't necessarily to build a bigger park next door. It might be to re-evaluate the current one. Is the factory land ratio too low? Is the road network taking up too much space? The researchers propose that there are specific, measurable rules for how to slice up the land to make it work. They aren't saying these rules are magic spells that will fix everything forever, but they provide a clear, data-driven starting point for planners.

In short, the paper tells us that building an industrial park is less like building a giant warehouse and more like cooking a specific dish. You can't just throw in a ton of ingredients and hope it tastes good. You need the right ratio of flour to sugar. If you put too much sugar (public space) or not enough flour (factory space), the cake fails, no matter how fancy the kitchen is. By finding these exact ratios, the researchers hope to help governments stop wasting land and start building parks that businesses actually want to fill up.