Last month, Gov. Arnold Schwarzenegger reignited a heated debate when he called for a civilized discussion on the merits of marijuana legalization. Indeed, the governor was responding to new public opinion polls showing greater interest in the policy idea -- and with the mounting problems associated with the drug trade in Mexico and here at home, it is hard to blame anyone for suggesting that we at least consider all potential policy solutions.

One major justification for legalization remains tempting: the money. Unfortunately, however, the financial costs of marijuana legalization would never outweigh its benefits. Yes, the marijuana market seems like an attractive target for taxation -- Abt Associates, a research firm, estimates that the industry is worth roughly $10 billion a year -- and California could certainly use a chunk of that cash to offset its budget woes in the current economic climate.

What is rarely discussed, however, is that the likely increase in marijuana prevalence resulting from legalization would probably increase the already high costs of marijuana use in society. Accidents would increase, healthcare costs would rise and productivity would suffer. Legal alcohol serves as a good example: The $8 billion in tax revenue generated from that widely used drug does little to offset the nearly $200 billion in social costs attributed to its use.

In fact, both of our two already legal drugs -- alcohol and tobacco -- offer chilling illustrations of how an open market fuels greater harms. They are cheap and easy to obtain. Commercialization glamorizes their use and furthers their social acceptance. High profits make aggressive marketing worthwhile for sellers. Addiction is simply the price of doing business.

Would marijuana use rise in a legal market for the drug? Admittedly, marijuana is not very difficult to obtain currently, but a legal market would make getting the drug that much easier. Tobacco and alcohol are used regularly by 30% and 65% of the population, respectively, while all illegal drugs combined are used by about 6% of Americans. In the Netherlands, where marijuana is de facto legalized, lifetime use "increased consistently and sharply" after this policy shift triggered commercialization, tripling among young adults, according to data analysis from the Rand Corp. We might expect a similar or worse result here in America's ad-driven culture.

An honest debate on marijuana policy also carefully considers the costs of our current approach. Arrest rates for marijuana are relatively high, reaching about 800,000 last year. Though these numbers are technically recorded under the category of "possession," the story that is seldom told is that hardly any of these possession arrests result in jail time (that is why former New York City Mayor Rudolph Giuliani made headlines when he aggressively arrested public marijuana users and detained them for 12 to 24 hours in the 1990s).

One of the most astute minds in the field of drug policy, Carnegie Mellon's Jonathan Caulkins, formerly the co-director of Rand's drug policy research center, found that more than 85% of people in prison for all drug-law violations were clearly involved in drug distribution, and that the records of most of the remaining prisoners had at least some suggestion of distribution involvement (many prisoners plea down from more serious charges to possession in exchange for information about the drug trade). Only about half a percent of the total prison population was there for marijuana possession, he found. He noted that this figure was consistent with other mainstream estimates but not with estimates from the Marijuana Policy Project (a legalization interest group), which, according to Caulkins, "naively ... assumes that all inmates convicted of possession were not involved in trafficking." Caulkins concluded that "an implication of the new figure is that marijuana decriminalization would have almost no impact on prison populations." This is not meant to imply that marijuana arrests do not have costs, but rather, that these concerns have been highly exaggerated.

Finally, legalizing marijuana would in no way ensure that the most vicious drug-related problems -- violence, economic-related crime, street gang activity -- would disappear. Most of those problems stem from the cocaine, heroin and methamphetamine markets. Marijuana's share of the black market is modest (the cocaine market is three times larger), and the money that is spent on the drug is spread over so many users and distributors that few are working with amounts that motivate or encourage high levels of crime.

Moving beyond the simplistic and unrealistic option of legalization, what can we do to reduce marijuana use and the costly harms it brings? Increasing the ferocity of enforcement isn't the answer, but increasing its potential for effectiveness through deterrent methods might be. Programs like Project HOPE in Hawaii, which perform regular, random drug testing on probationers and others and implement reliable, swift (but short) sanctions for positive screens, have shown remarkable success. Innovative solutions, grounded in sound research on prevention, treatment and enforcement, present the shortest route out of marijuana-related costs. But an open market for the stuff? That doesn't pass the giggle test.

Kevin A. Sabet worked at the Office of National Drug Control Policy in the Clinton and Bush administrations. He is currently a consultant in private practice.