A Brilliant Used Car Salesman? How Shift Solved the ‘Agency Problem’
Freakonomics + The Principal-Agent Problem
I love Freakonomics. I loved the book when I read it a long time ago, and now I get my fill by being a pretty avid listener of the podcast (plug for Stitcher as my go-to podcasting app). They’ve really nailed the ratio of accessibility to nerding out.
One of the lessons of the book that most stuck with me was the misaligned incentives between real estate agents and home sellers. In short, it is a classic ‘principal-agent problem’, in which one person or entity (the “agent”) is able to make decisions on behalf of, or that impact, another person or entity (the “principal”) and the incentives of those two parties are not necessarily well aligned. As research has shown and Freakonomics helped make widely understood, real estate agents have little incentive to help home sellers actually get the best price for their home. More info (here, here, and here), but basically getting an extra $10,000 for your house means a lot to you as the seller and very little to them as the real estate agent (based on the small cut of that $10,000 they’d pocket). They are incentived to get you to sell quickly and move on to the next sales opportunity; they still get 90+% of the value and don’t have to put in the extra work to help you squeeze out the most possible value.
Proper alignment of incentives must be one of the most important lessons (and pitfalls) in business and life, and this example has always stuck with me. It came back to me recently when looking at Shift — a marketplace for buying and selling used cars — and their fairly ingenious pricing model that seems to have overcome the principal-agent problem.
Shift’s Pricing Model
Below is the summary of Shift’s approach to pricing from their website. If you decide to sell your car with Shift (rather than going through the hassle of dealing with Craigslist buyers or getting ripped off by trading in to a dealership), the process entails:
- After inspecting your car, Shift gives you a guaranteed price for your car; no matter what, you’ll receive this amount of money. The seller can agreed to this price or decide not to move forward with Shift.
- Shift actually sets the price your car will be listed at on their marketplace; using their data and insights about car buying in your area for similar cars, you’d expect they’d be able to better estimate the ‘right’ price.
- Shift takes 50% of the final sale price above the guaranteed price it provided it step 1.
I think this is a fairly ingenious way to get around the agency problem. It fully aligns Shift’s incentives with the seller’s. Shift is motivated to get as much as possible for your car (including by detailing your car and doing some minor repairs), since their cut is contingent upon getting more than the guaranteed amount. What they’ve effectively done is say:
- You could sell your car for the guaranteed price today without much hassle (i.e., by putting it on Craigslist and there’s likely to be a buyer at that price)
- We can help you get more than that guaranteed amount (and take a lot of the work off your hands). We’ll only get paid on this additional upside (and we’ll take half of any value we create).
A Better Model for Residential Real Estate?
Imagine if selling your home worked like this. The home seller interviews real estate agents, and each agent offers the seller a ‘guaranteed price’ for their house; the agent won’t make a dime unless the house sells for more than that amount. The seller chooses the agent + ‘guaranteed price’ that they like the best. The agent then makes a big cut of any amount the house goes for about that guarantee.
Would the math work? Let’s use an oversimplified example of a house that might sell for around $500K:
- Today, on a $500K sale the agent’s cut would be typically be around 6% or $30K (which typically gets split with the buyer’s agent)
- Now imagine the agent had agreed to a guaranteed price of $450K and then a 50% cut of any amount above that. If the house went for $500K, the agent’s cut would be $25K (50% of $50K). If they put in the extra work at got $525K for the house, that cut would jump to $37.5K, and all parties would be better off!
I don’t know if this is how real estate will work in the future, but it seems unlikely we’ll stick with the current model. 1 vote for Dubner and Levitt to take a shot at fixing it.
