When Republican presidential candidate Rick Santorum suspended his presidential campaign last month, the former Pennsylvania senator all but sealed Mitt Romney’s easy victory in the state’s April 24 primary.

Santorum also dashed the expectations of his home state’s broadcasters, who were counting on the candidate to keep the race competitive and their ad inventory—much of which had already been reserved by Romney’s campaign—in high demand.

The day after Santorum dropped out, The New York Times reported that “Romney aides immediately went to work canceling what they expected would be a $2.9 million advertising campaign in Pennsylvania, a huge savings equivalent to roughly 40 percent of the cash Mr. Romney had on hand at the end of February.”

The development led Kathy Kiely of the Sunlight Foundation to claim on the organization’s blog that the “biggest loser” in Pennsylvania primary was not Santorum, but the state’s 46 local broadcasters.

That got us thinking: we’re sympathetic to the argument that the windfall that local TV outlets gather during the political season ought to be returned, to some degree, to the public in the form of stronger political reporting.

But, given that there are only so many hours in which to sell ads—and other potential advertisers ready to buy that inventory—just how big is that windfall? And just who are its beneficiaries?

The answer, as with many things: it depends.

An ‘unprecedented frenzy,’ unevenly spread

Projections vary, but all observers agree that the 2012 campaign is likely to set records for TV ad spending. Citing a “near-perfect political storm,” a Moody’s Investors Service report issued last June predicted an “unprecedented frenzy of political advertising” and 9-18% revenue growth above 2010’s $2.3 billion, the current record. During an event for media buyers in New York last month, Ken Goldstein, president of Kantar Media’s Campaign Media Analysis Group, offered a higher ceiling, estimating $2.5-$3.3 billion will be directed to local spot advertising.

Those projections are based on the assumption that local television will attract a little less than half of all campaign dollars, as it has in the past. That assumption is likely to hold up, observers say. Even as campaigns diversify their messaging strategies and develop new approaches to the ground game, television remains the dominant way to reach potential customers—and swing voters.

But the distribution of those dollars is far from even. The bulk of the cash, of course, flows to stations in politically significant markets—presidential battleground states, or those with early primaries, or especially competitive Senate races.

Even within those markets, though, ad buys are heavily skewed toward top-ranked stations. Moody’s analyst Carl Salas says a market’s top broadcaster will typically draw 40-65% of the political spending, the second-rated draws about 30-40%, and the bottom two stations will split the balance, perhaps 5%. And while strong network programming boosts ratings and attracts ads, the greatest value lies in having the top local newscast, as the station owns that ad inventory and gets to keep 100% of the revenue it generates.

Just as the ads are concentrated by geography and station, they are also compressed by time. An analysis of 2008 and 2010 by a TV industry trade association found that political ad share builds from roughly 5% of all inventory around Labor Day to 20-25% in the days before the election. The shares are larger in battleground states. In the Harrisburg market, political spending accounted for 32% of spots and 48% of advertising revenue in the week before Election Day; in the Columbus market, political buys made up 50% of spots and 61% of revenue for that week.

All that concentrated demand puts pressure on prices to rise, but that doesn’t mean campaigns are paying top dollar. By law, candidates for federal office are entitled to rates that are the “lowest unit charge per class of time” within 45 days before a primary, or 60 days before a general election. That means they receive the discounted rates given to loyal, favored advertisers. (While the campaigns get the lowest rates in a given ad class, they do tend not to buy the cheapest class, which is subject to “immediate pre-emption.”) In fact, Goldstein and his Kantar colleague Elizabeth Wilner argue that the primary motivation behind broadcasters’ opposition to a new FCC rule mandating that political ad buys be posted online is that the broadcasters don’t want ordinary non-political advertisers to know just how low those rates are.

Super PACs and issue groups, though, are not entitled to the same low rates. And in a post-Citizens United world, those groups account for an increasing share of political ad buys. In 2002, candidates accounted for about 60% of political ads; in 2010, it was down to about 40%, and it’s expected to fall further this year. And as the campaigns buy up ad time, and drive up costs on the remaining inventory, the rates for super PACs—and other advertisers who just have to have that 30-second spot in that hour on that day—shoot up. The difference between the campaign’s “lowest unit rate” and what a super PAC might pay can be significant, says Goldstein—anywhere from 150 to 5000 percent.

Put all the right conditions together, and the effect on ad prices can be extraordinary. According to Goldstein, a spot in Des Moines, Iowa, on Dec. 30—in the middle of the Insight Bowl, which pitted the University of Iowa football against Oklahoma; and four days before the Iowa caucus—went for $15,000. The cost normally would have been $500.

‘You can’t count on political money until the check arrives’

The targeted nature of political ad buys sorts broadcasters into groups of big, big winners and (relative) losers. As the targeting becomes ever more fine-tuned, though, even the winners have to accept some unpredictability.

That instability is driven by polling, which, as it has become finer science and bigger business, is giving campaigns and super PACs more precise and time-sensitive information about which voters to target.

Bob Prather, the president of Gray Television, an Atlanta-based company which operates 36 stations in 30 small to medium-sized markets, said this effect led Republicans in 2008 to pull $800,000 in ads from his Ohio stations as they focused their efforts elsewhere. Three days later, Democrats cancelled their $500,000 buy.

“The thing I’ve seen in 20 years is that the spending is a lot more sophisticated and controlled,” said Prather. “They move money around real fast depending on polling. It seems to be more volatile every election cycle.”

In some cases, spending is also compressed into shorter time periods. That was certainly the situation this year for WIS TV in Columbia, S.C., at the center of a key early primary state. Scott Sanders, the station’s general manager, said just $100,000 had been spent in his market as of Jan. 3; by the Jan. 21 primary, the figure had ballooned to nearly $2 million.

“Four years ago, candidates started advertising in October of 2007,” said Sanders. “This time, it started two weeks before the primary and was lumped into one big sum. Super PACs came with basically unlimited money and said, ‘We’ve got X amount to spend, where can you put it?’ For a 12-day period in January it was unlike anything I’d ever seen.” (Sanders said he actually exhausted his inventory and had to turn some super PACs down, just the doomsday scenario pondered in a recent Campaigns & Elections article headlined, “Could we run out of airtime?”)

In other words, it was the flip side of what happened in Pennsylvania this year. But Holly Steuart, vice president and general manager of a CBS and CW affiliate in the Harrisburg market, was philosophical about missing out on that presidential primary spending.

“You can’t count on political money until the check arrives,” said Steuart. “Romney is a good example of that—this spending is very unpredictable. It could amount to nothing, or it could be beyond your wildest dreams.”

Political peaks and ‘hammock years’

Of course, it’s easier to be philosophical when there are other campaigns to keep the cash coming in. CJR has already noted how stations in Scranton and nearby markets benefited from a contested Democratic congressional primary that drew in substantial spending from outside groups. Pennsylvania’s Republican Senate primary was also good to broadcasters, with the victor, Tom Smith, spending $6 million of his own money. (Station heads tend to have a nonpartisan appreciation for the candidacies of the 1%. Gray’s Prather recalled with fondness Bruce Lunsford, who sunk $8 million of his own money into Kentucky’s gubernatorial primary in 2003 only to drop out several days before the vote.) Steuart said her station lost several thousand dollars worth of Romney’s ads, but the cancellation freed up space for “healthy spending” in competitive local and state-wide races, and also for commercial advertisers.

That brings us back to one of the questions we started with: To what extent are the political ad buys actually producing “new money” for the stations, and how much are they just taking the place of commercial ads?

Campaign ads are “not all just gravy to the stations,” said Steve Lake, the senior manager of national accounts at SourceMedia, which owns KCRG, the dominant station in the Cedar Rapids, Iowa. “There is lost revenue associated with political advertising. Try as we may, sometimes we cannot make good on all the spots that get displaced because of time-sensitive nature of events. There is an amount where you say we’re just going to have to credit it.”

In other words, those Pennsylvania broadcasters were probably not a full $2.9 million poorer after the Romney campaign canceled its buy—and likely wouldn’t have been even without other campaigns picking up the slack.

Still, most of the time, campaign ad dollars are mostly gravy. That’s partly because the concentrated demand drives up prices for all ad buyers. But it’s also because stations often don’t have to turn away commercial ads to accept political buys. Advertisers buy inventory in packages—they may, for example, buy a certain number of ads to run within the month, or within the quarter during a certain “daypart” (industry-speak for time of day). That allows broadcasters a fair amount of flexibility to juggle inventory and make way for time-sensitive ads that need to air during a specific day or hour.

Then there’s the fact that ad inventory is not exactly finite. Salas, the Moody’s analyst, explained that local broadcasters can accommodate high demand by creating more news programs and thus more high-value advertising spots. More common is for stations to temporarily add or lengthen news programs, or simply to add more spots into existing programming.

One way to get a handle on just how much stations benefit from political ads is to take a closer look at Gray Television, an industry leader in terms of the share of revenue generated from campaign buys. During the last mid-term elections, political ads accounted for 17% of the company’s total revenue, nearly double the industry average of 9%.

In 2010, that leading position earned the group a record $57.6 million from political advertising. The following year was also a record for a non-election year, but political revenues still fell by over three-quarters, to $13.5 million. That’s a reduction of $44 million in political revenues—which nearly matches the company’s $41 million overall decline in ad revenues. In other words, as political buys plummeted, revenue from other ads remained mostly stable.

That’s an imperfect calculation that obscures other areas of growth and loss; in an interview, Prather, the company’s president, said revenue sources vary too much from market to market to come up with a general rule.

But for those stations that win the campaign ad sweepstakes, election years do offer a reliable boost. Even KCRG’s Lake, who said the ads are not “just gravy” for stations, had an evocative phrase for the cycle: election-year peaks help to cover the less lucrative “hammock years” in between, with the added revenue going toward capital investment (or, for many stations, helping to pay down debt).

That volatility and unpredictability is something the stations can learn live with, given the money at stake. Prather said Gray has become a favorite for political buyers by fortunate accident—the company’s strategy has been to invest in leading stations in university towns and state capitols, which tend to have stable economies, rather than in battleground states per se.

But he’s not regretting the result. “We think it’s a good business,” Prather said. And he’s keenly aware of quirks in the calendar that make business even better. “The best thing that is the election is on November 6 this year,” he told me. “Those six days we’ll get a huge amount of political—if it’s on November 1 you miss all those days.”

In an election year like this one, are there any drawbacks? Maybe only that the other years seem less sweet.

Erika Fry is a former assistant editor at CJR.