A real-estate agent keeps her own home on the market an average of ten days longer [than she would for a client] and sells it for an extra 3-plus percent, or $10,000 on a $300,000 house. When she sells her own house, an agent holds out for the best offer; when she sells yours, she encourages you to take the first decent offer that comes along. (more)
Lecturing on incentives on Wednesday, I used the classic example of the bad incentives of real estate agents. They usually get a fixed percentage (3%) of the sale price, which mostly makes them want to close a deal as fast as possible, regardless of the sale price. This is bad for seller’s agents and positively perverse for buyer agents – they worse the deal they get for you, the more they get paid. And the scope for individual agent reputations is pretty limited, because most people only ever buy or sell a few houses in their lifetime, usually in geographically separated places.
Alex just posted on the continuing puzzle of why this fixed percentage doesn’t seem to respond to changing market conditions, arguing that neither monopoly nor signaling explains it, and suggests:
Part of the problem in the realtor market is that other realtors can easily discriminate against discount brokers by pushing their clients one way or the other. (more)
That may be why we won’t see something better soon, but my lecture prompted me to think about the still interesting question: what exactly would be a better contract between you and your real estate agent, one that would better align their interests with yours?
Searching I found this paper from 2000, which proposes that selling homeowners sell their home to the selling agent, but also give that agent an option to sell the home back at the same price, to give that homeowner an incentive to help sell. They make no suggestion about how to contract with a buyers agent.
Here is what I came up with after my lecture. On the sell side, have the homeowner sell a 20% stake in their house to the selling agent, for an agreed-on cash price. The homeowner might hold an auction to find the local agent willing to pay the highest price to take on this role. The agent turns that back into cash when the house actually sells, or if it doesn’t sell the agent can sell their 20% stake back for the same price they paid if they want to give up on the process for now. If the homeowner wants to give up on the process, a similar reverse sale would happen, but perhaps the homeowner should suffer a penalty, such as 10% of that price paid for the 20% stake.
On the buy side, I’d have the buyer agent agree to pay (20-X)% of the house purchase price to gain a 20% stake in the house at the time of the home purchase. The X% number would be the agent’s fee, which might be chosen by an auction among the local agents. Unless they could find someone else who agreed to buy this stake after the purchase, they’d have to hold on to it until the house is next sold. Perhaps for many years. Because the buyer would get to live in the house or rent it, while the agent would not, the homeowner would owe the agent 20% of some assigned rental price each month until the house was again sold. This rental price could come from a simple regression of rental prices on local home features. People would know this price wasn’t exactly right, but they could take deviations into account in setting the price X.
The 20% number in the above is obviously arbitrary. It is probably a better place to start than the 100% number in the other proposal I mentioned, being a less radical change from the status quo. But my proposal is really to have some percentage number like that, not the exact 20% number.
This proposal clearly requires the agents to take on more financial risk than they do today, and so would encourage them to organize into agent firms that jointly take on the risk together. But that seems pretty reasonable.