There’s been some buzz about the FCC’s latest draft order on spectrum policy, which updates rules for the new Citizens Broadband Radio Service (CBRS) in the 3.5 GHz band. Some critics have called this item a “gutting” of their preferred outcomes. Such reactions are unwarranted. The new order is actually a substantial improvement over the older proposal.

The updated CBRS proceeding would allow the private sector to access parts of the spectrum that to this point have been reserved for naval radar. Because the Navy isn’t using the spectrum everywhere, all the time, the plan is to maintain the Navy’s priority in the 3.5 GHz band, but allow other parties to use it when they are not.

The plan divides the band into 15, 10 MHz-wide channels. Eight will be General Authorized Access (GAA), which is akin to unlicensed frequencies that anyone can use as long as their device meets certain specifications. The remaining seven channels will be auctioned off by the FCC as Priority Access Licenses (PALs)–although they can be used opportunistically by GAA devices when not otherwise in use.

Much of the current controversy concerns the size of the license area for those PALs. The original proposal was for the area covered by each PAL to be census tracts, which are roughly the size of a few city blocks. The new order would increase the license area to correspond with counties, resulting in fewer, much larger PALs.

Critics of the move argue that it tears the band away from potentially innovative new business uses. In their vision, hotels, stadiums, factories or other establishments would buy PALs to create networks for their own unique needs, rather than buying traditional wireless services from a carrier, like Verizon or AT&T. But this vision only works, they claim, if the PALs are small enough that businesses can afford them. A rational hotel owner won’t buy a license for the whole county just to operate a network in one building.

This argument is flawed, however, and the new FCC order with county-sized PALs is likely even better for individual venues than the old one. The argument against county-sized PALs is that they make businesses buy larger licenses than they need, which could prevent them from buying a license at all. The problem is that census tracts are themselves much larger than individual businesses. For example, here in Washington, D.C., census tract 49.02 contains the Washington Convention Center, the Marriott Marquis hotel and several other businesses and apartment buildings further north. If the hotel wanted to buy a PAL for that census tract it would still have to pay for a much larger area than it needs.

In short, regardless of the original size of PALs, they will never be the right size for every user. But, as we explained in our comments to the FCC, the new order fixes this problem by allowing PAL owners to sell or lease portions of their license. This way, the hotel can go to the owner of a county-sized PAL and buy exactly the size it needs. It’s akin to the age-old scandal of hot dogs being sold in packs of six while hot dog buns are sold in packs of eight: it would be better for one person to buy buns in bulk and sell hot dog customers exactly the number they need.

Critics downplay the possibility of secondary market transactions such as these on the grounds that smaller businesses don’t often get much of the spectrum on secondary markets. But this criticism is misplaced. The fact that secondary markets generate outcomes different from those that certain advocates or businesses prefer does not mean they don’t work.

Indeed, if our example hotel doesn’t come away from the secondary market with a section of a PAL for its business, we should expect that they wouldn’t have acquired one even if the PAL areas were census tracts. If census-tract advocates had gotten their way, the hotel would still have to bid against the larger carriers for a PAL. And if the carrier is willing to turn down offers to lease a small portion of a PAL in the secondary market, the same carrier would likely outbid smaller companies for the whole census tract. The two scenarios are economically equivalent: in both cases the carrier proves that it would rather have a PAL covering the area in question than the revenue it could acquire or save by forgoing such a license.

In the end, the initial size of licenses is not that important, because the size can be customized in secondary markets. However, this theory could obviously apply to a parallel argument that larger licenses should be created out of many smaller ones. Which size is best, therefore, must depend on other factors, like the productivity of the band or the efficiency of secondary markets.

By these two metrics, larger licenses are probably better. Since radio waves cannot be made to stop exactly at the boundary of a license area, licensees often have to reduce power levels or risk crossing into someone else’s turf. This means that the band cannot be used to its full potential in those regions. The more of those boundaries there are, the worse the problem gets. And since smaller licenses necessarily entail more boundaries, larger license areas are more efficient.

We must also examine the transaction costs associated with using secondary markets to break down larger licenses or put together smaller ones. Here again, it seems likely that larger, more experienced licensees, who can profit from leasing portions of their PALs to parties that highly value a certain portion, will be better equipped to facilitate these transactions than parties that have very little experience owning and operating spectrum licenses.

Far from “gutting” CBRS, switching to larger PALs will be beneficial to the overall productivity of the band and will increase the likelihood that this innovative proceeding will be successful.