It was a bit more subtle than that.
Given a fixed percentage (let's take the 6%), if the agent sells today the house (valued 100,000-120,000 US$) he/she gets US$ 6,000 today.
If the 100,000 offer is refused the house stays on the market for - say - 3 months more, and eventually is sold at 110,000 US$) the agents gets only US$ 600 more (and three months later), but has to arrange many more visits, so the formula is incentivating "quick sales" and the agent has all the interest to counsel the seller to accept a "low offer".
The proposal by the freakanomics authors was to increase the percentage due to the agent on the sums exceeding the "base value" of the property, so that the convenience would be shifted to "sell the house at the highest reasonable price in a reasonable time".
Real estate agents are more likely to plan sufficient time to maximize sale value; most people make plans with timelines where sale (or purchase; I bet you’ll find a similar issue on the other side, too) of the house is a blocker for other things before talking to an agent, so the disruption of not completing the transaction timely is significant.
The reasoning they gave was that the difference in the realtor's percentage between a lower total price and a higher total price was not meaningful enough for them to do extra work or take extra time to get it. It is better for them to sell more houses for less, rather than fewer houses for more.
So yeah, the percent means the incentives are more aligned, but maybe not aligned enough.