The housing market seems to be slowing down. Despite the vast appetite, the shortage of inventory and the historically high-interest rates are forcing those looking to buy their forever home to postpone their dreams. Even before the interest rates reached the current level, buying a house was still very challenging. In some markets, the competition had become intense (in states like Massachusetts). Before we listed our house, there was an average of 8 offers on any listed house (though somehow that did not happen to us for some weird reason).

One of the reasons for this intense competition is the lack of inventory. Based on my observation while I was in Massachusetts and on observations made now living in a Chicago suburb (which is so densely populated that you can see what your neighbor is eating for dinner), a reason behind this shortage is archaic zonings.

Zoning laws in most U.S. states were defined decades ago with good intent. But like every good thing, their relevance needs to be revisited. Not that they are not relevant. What needs to be analyzed if the zoning method is still relevant. You can drive for miles and miles in many Boston area suburbs with a competitiveness score of 95+ on Redfin and still see only a few houses. There are stretches of land zones as commercial, with no development, or strip malls with minimal occupancy.

The challenge of this intensely competitive market has implications beyond the housing market. Young couples will relocate if they see no realistic possibility of affording a house in a region. No matter how lucrative industries and jobs you currently may have, the nature of work and distributed physical presence of companies across the U.S allows young people to look for more affordable areas with good job prospects. There is a slow and gradual negative impact on the economy.

How can analytics help, though?

By helping analyze the current state of zoning and optimize a future state while meeting needed criteria. Zonings, at least in Massachusetts, is defined by local ordinances. With hundreds of thousands of documents defining zones across the state, it is humanly impossible to analyze all of them and interpret and summarize the criteria. This is the first step where a deep learning algorithm can help categorize and summarize the details of every zoning criterion.

The second step is to understand the relevance. This is where human insights need to be leveraged as well. AI algorithm can help categorize every zoning criterion across all ordinances into a few specific categories. Humans then need to review the categories based on the current state scenarios. The algorithm can leverage plenty of data points to insert its own comments in these categorizations by ordinance. For example, suppose a village allocated a lot of parcels for commercial purposes because it felt the need for that in the 1990s. In that case, the current data points, like businesses that have opened and then closed in those establishments, current occupancy, financial performance, population density, and trends in that subdivision, can be leveraged to evaluate if those zoning criteria still hold.

There is no end to fine-tuning such an algorithm. Geoanalytics can be leveraged as well. If a parcel of the land was flood-prone a few decades ago, what does it look like now? Can the land be reclaimed from nature with some investment?

The pace at which the housing crisis is growing in some markets could be a great business opportunity for consulting companies. An algorithm paired with the abovementioned analysis and many states should jump at this service. After all, the long-term health of their economy may be at stake.