ChatGPT failed the housing test…
Many folks assume that rising interest rates crash the housing market. After all, higher rates mean buyers need to pay more for the same house over the length of their loans.
But last Thursday, my colleague Briton Hill showed how that assumption is wrong…
In short, the federal government enacted several programs to support housing through the rising-rate period of the 1970s. And as he noted, it’s starting to do the same thing today.
We’ve also written time and again about the high demand for housing in the U.S. And we’ve reasoned that the best way to meet homebuyers’ needs is through increased supply.
So it’s clear that we’re “bullish” on homebuilders at Chaikin Analytics right now. And with that in mind, I recently asked ChatGPT a simple question…
Will high interest rates hurt housing?
As you’ll see today, ChatGPT worked great as what I call a “trade-show tool.” These tools impress crowds of people who don’t take time to dig deeper into what they’re seeing.
But beyond that, it failed miserably…
Like a trade-show tool, ChatGPT sounded like an expert at first. Its response began with…
High interest rates can potentially have an impact on the housing market, although the extent of the impact can vary depending on several factors. Here are some ways in which high interest rates can affect the housing market.
Then, it listed four factors to explain how high interest rates can hurt housing.
The trade-show experience continued with the first factor. It’s the most obvious response…
High mortgage rates reduce the pool of buyers who qualify for financing.
Next, ChatGPT said housing demand would fall. Potential buyers might decide to not go forward. (That sounds similar to the first factor. But it’s a computer, so I’ll be lenient.)
Then, it discussed how homebuilders would find it harder to finance their own operations.
And finally, it said higher rates hurt people who have adjustable-rate loans and need to refinance. In that type of environment, it increases the risk of foreclosure.
That all sounds smart enough. Again, ChatGPT would dominate the trade-show circuit.
But ChatGPT lost me with its grand wrap-up…
It’s important to note that while higher interest rates can have an impact on the housing market, they are not the only factor influencing its dynamics.
Other factors such as local market conditions, supply and demand dynamics, employment levels, and overall economic conditions also play significant roles.
Real estate markets can be complex and are influenced by a multitude of factors, so it’s advisable to consider a comprehensive range of factors and consult with experts when assessing the potential effects of interest rates on the housing market.
That’s not artificial intelligence (“AI”). It’s just artificial. It lacks any hint of intelligence.
The first two paragraphs vaguely regurgitate and touch on the original question…
They read like a lawyer wrote them, expecting someone to sue over a bad answer. They’re vague enough that ChatGPT’s defense could be, “Wait, we didn’t really give that conclusion.”
It’s an AI hedge against being wrong. And in reality, ChatGPT didn’t give any conclusion!
That’s my point.
And the final paragraph doesn’t add anything to the rest of the response. It’s just a bunch of big words attempting to make the reader think it’s really smart – like a trade-show tool.
If ChatGPT were a truly intelligent resource, it would’ve dove deeper into the “other factors” instead of just quickly listing them. It would’ve broken everything down for novice followers.
In the end, ChatGPT doesn’t yet think and act like a human expert. It’s still a work in progress.
ChatGPT and similar AI tools will likely evolve in the months and years ahead. They could potentially become more than just trade-show tools at some point in the future.
But as you can see, we’re not there yet.
Good investing,
Marc Gerstein
