It's tough being corrupt. You know you probably need to bribe lots of people to achieve your ignoble goals. But where, exactly, do you put your money? Do you invest more heavily in payoffs to politicians, to the police or to judges? Or would it be more effective to bribe the news media to ignore scandals and produce monotonously favorable coverage?

The answer is to invest in the news media, claim economist John McMillan of the Stanford University's Graduate School of Business and graduate student Pablo Zoido in a paper to be published in the fall by the Journal of Economic Perspectives.

They based their conclusion on a detailed study of an unusual data set: Records of bribes kept by Vladimiro Montesinos, the former head of Peru's secret service. Montesinos's goal was to protect his patron, then-President Alberto Fujimori. At the height of the spy chief's dirty dealing in the 1990s, he was dishing out payoffs totaling more than $3 million a month to police officials, key judges, opposition political leaders and the owners of the country's major television stations, according to records.

The money came from the public treasury. Most of the bribes were paid in U.S. dollars rather than Peruvian soles. "On one video Montesinos apologizes to the bribee for paying in soles, as he didn't have dollars on hand," McMillan said in an e-mail.

By virtue of their sheer volume and detail, the bribes and tape collection are noteworthy -- as well as a rich source of information for researchers interested in the economics of bribery.

"Montesinos kept meticulous records of his transactions," Zoido and McMillan recounted in their paper. "He required those he bribed to sign contracts detailing their obligations to him. He demanded written receipts for the bribes." And if that weren't enough to secure him a place of honor in the Hall of Shame, he "had his illicit negotiations videotaped."

That allowed McMillan and Zoido to tally up the bribes by institution and then compare them to see where Montesinos, in his "expert" opinion, thought it most effective to spend the most money.

It wasn't even close. "One single television channel's bribe was four times larger than the total of the opposition politicians' bribes," they found. "By revealed preference, the strongest check on the government's power was the news media."

The totals were staggering. Bribes to the owners of the six largest privately held television stations amounted to millions, with $1.5 million a month flowing to the owner of one station and $500,000 a month to another, while a third got a total of $9 million over an undisclosed period of time. Sometimes the bribes were payments for specific jobs: Montesinos paid the owner of one TV network $50,000 to fire two reporters critical of the government.

By comparison, politicians and judges came relatively cheap. Bribes to 21 judges came in the form of one-time payments that ranged from $2,500 to $55,000. Lawmakers were a bit more pricey: Dozens of opposition party legislators were paid bribes ranging from $10,000 to $50,000 a month, easily outstripping their monthly paychecks.

Newspapers were less of a concern because the overwhelming majority of Peruvians got their news from television, McMillan said. Still, Montesinos paid more than $2 million to the owners or directors of popular newspapers.

The implications of the economists' findings extend far beyond Peru. These findings suggest that if a society wants to prevent corruption, the best weapon is an independent and corruption-free media that can act as a check on the government.

It was poetic justice that Montesinos's videotaping habit ultimately proved to be his undoing. One of Montesinos's boodle videotapes -- showing him paying off an opposition party leader -- was broadcast in September 2000 by Peru's single untainted TV station.

Other tapes quickly surfaced. They soon became "Peru's own distinctive form of reality television," the researchers noted. Fujimori shortly fled to Japan, where he was granted asylum because his parents were Japanese-born. He resigned the presidency by fax.

Montesinos also fled the country but was arrested in Venezuela and faces multiple charges of corruption, drug trafficking, arms dealing and human rights abuses. Last month, he was sentenced to 15 years in prison for paying television station owners millions of dollars to broadcast favorable news about Fujimori.

Alcoholic-Enablers Anonymous?

Hey, where can a falling-down drunken fella buy a drink around here? Sadly, the answer appears to be just about anywhere.

At least that's what researchers at the University of Minnesota found when they tested the willingness of bartenders and store clerks to sell alcohol to patrons who clearly appeared to be three sheets to the wind.

Traci L. Toomey, an associate professor of epidemiology, and her colleagues hired actors to impersonate drunks and then sent them to 355 liquor stores, bars and restaurants in and around Minneapolis to try to buy a drink or a six-pack of beer. Observers were present at each locale to witness the attempted transaction.

The buyers put on Oscar-worthy performances -- staggering, stumbling into things, slurring their speech and showing other signs of acute inebriation. No matter. They were sold alcohol in 76 percent of the bars and 83 percent of the stores. (Minnesota and other states outlaw the sale of alcohol to people who are obviously drunk.) Even when sellers clearly noticed the drunken behavior, they sold alcohol to the phony customers 61 percent of the time, with younger clerks and barkeeps far more willing to make the sale than older ones, Toomey reported in a recent issue of the journal Alcoholism: Clinical and Experimental Research.

Breakthrough Breakdown

There's a dark side to the recent good news about breakthroughs in the campaign against HIV, the virus that causes AIDS: Receiving treatment with the anti-retroviral drug "cocktail" designed to delay the onset of the disease seems to produce "more sexual risk-taking by HIV-positive adults and possibly more of other risky behaviors like drug abuse," reported three researchers from the Rand Corp. in a recent paper published by the National Bureau of Economic Research.

They based their conclusions on three monitoring surveys of 1,396 HIV-positive men conducted between March 1996 and January 1998, a time when the three- or four-drug cocktail was becoming widely used. They found that receiving anti-retroviral treatments "reduces the probability of having no new sex partners by 39 percentage points and increases the probability of having two or more partners by 33 percentage points," wrote Rand researchers Dana Goldman, Darius Lakdawalla and Neeraj Sood.

They said their findings may help explain this puzzling pattern: The AIDS death rate and the rate of HIV incidence fell significantly between 1995 and 1998, but then the infection rate started rising while the death rate continued to drop.