The violent crime wave of the late 1980s and early 1990s is a subject of perennial debate among policymakers and social scientists: what caused the spike in America’s murder rate, which by 1993 soared to the highest level ever recorded? And just as important, why did it subside?

One popular theory attributes both the surge and ebb of homicides to crack cocaine. The logic is persuasive. In poor urban areas hollowed out by deindustrialization and cut off from economic opportunity by racial discrimination, the drug provided one of few lucrative incomes for young black men. Dealers resorted to bloodshed to defend their businesses, while users turned to crime to feed their addictions. Then came the “war on crime” and go-go economy of the Clinton years, which dried up the crack trade and reduced murders along with it.

But two new academic papers posit that drug market dynamics alone don’t fully explain why the explosion of crack use was so deadly, nor why murders fell in the mid-90s. Instead, they argue, a boom in handgun production and possession gave the crack years their fatal character — until new restrictions on firearms reversed the trendlines.

“What’s striking about the gun market is you get these surges in production,” said Geoffrey Williams, an economist at Transylvania University in Kentucky who has been researching the phenomena for the past three years. “The production booms were followed by surges in killings.”

In a working paper updated in August, Geoffrey Williams and his colleague W. Alan Bartley argued that it was a “supply shock” of low-priced pistols which in the 1980s and early ‘90s led to higher levels of gun homicide among young black men. During those years, ATF oversight of the gun industry slacked off and a group of Los Angeles-based manufacturers known as the “Ring of Fire” expanded the market for “Saturday Night Specials”: bottom-of-the-barrel firearms that turned up at crime scenes as surely as weekends beget drunken disputes.

Culling advertisements in back issues of the magazine Gun Digest, Williams and Bartley compiled handgun price and production data for the period. They calculate that the production of cheap guns priced at $100 or less peaked in 1993, the same year murders, both those committed with guns or other means, reached their highest point. That year, Washington, D.C., had a murder rate of 75 per 100,000 residents. The industrial hub of Gary, Indiana, led the country at a rate of 110 per 100,000 residents. Today, very few cities even approach those levels of violent death.

The huge pool of cheap guns contracted sharply as the 1990s wore on. Manufacturers of cut-rate handguns were driven out of business by product liability lawsuits (their guns tended to injure users, too). The federal government doubled ATF law enforcement funding, from just over $2 billion in 1990 to more than $4 billion by 1994. The then-nascent Brady background check system, which allowed gun dealers to instantly check whether a purchaser was prohibited from owning a gun, reduced the ability of gun purchasers with felony histories or other disqualifying behavior to buy firearms.

By 2000, the cheap handguns that had risen to claim the largest share of production sunk to the smallest share. Over the same years, the gun homicide rate among young black men also fell.

So did their suicide rate, which Williams and Bartley consider crucial evidence. By law, there is no government count of private gun possession, so scientists who’ve sought to study the effects of gun ownership have needed to develop a proxy measure. Several broadly accurate substitutes for direct counts of firearm ownership rely on a state’s rates of gun suicides. (See this Rand Corporation write-up for a detailed explanation.) Williams and Bartley, borrowing that metric, see the decline in suicides among young black men not only as an indication that fewer were fatally shooting themselves, but also that fewer had guns at all.

The professors found what they take as a further counterpoint when they plumbed data relating to drug use. If the crack boom propelled murder rates, as that theory holds, then falling homicides should have trailed a decline in rock cocaine use. But the opposite was happening: Data they pulled from the National Bureau for Economic Research suggest that crack use actually continued to increase after 1993. (Some scholars believe it did not peak until 2006.) Here, the number of cocaine overdoses provided the proxy: since crack is smoked, it enters the bloodstream faster than powder cocaine, presenting a higher risk of overdose. “Cocaine overdoses almost never happened before crack,” said Williams, making those deaths one way to track the crack market.

Crack was associated with a rise in one form of crime, Williams and Bartley found, but according to their research, it wasn’t murder — it was property crime. What yielded more killings was the guns that flooded city streets around the same time.

Where Williams and Bartley draw a link between gun production and gun possession and skyrocketing homicide rates, a separate working paper suggests the effects of the increased demand for guns instilled by the violent crime of past decades persists today.

In findings published by the National Bureau of Economic Research in July, economists William Evans of the University of Notre Dame, Craig Garthwaite of Northwestern University’s Kellogg School of Management, and Timothy Moore of Purdue University look at a terrible exception to the crime decline of the past two-plus decades. While killings ebbed from their early 90s highs for nearly every demographic, the murder rate for young black men remains 25 percent higher than it was before the crack epidemic.

The crime wave of the 90s passed, and “it’s easy to think that it’s gone,” said Moore. “But younger black males are still doing worse than other groups.”

According to Evans, Garthwaite, and Moore, the reason homicide rates for young black men remain high is the lasting effects wrought by increased access to and demand for firearms during the crack years.

“The diffusion of guns both as a part of, and in response to, these violent crack markets,” the authors write, “permanently changed the young black males’ rates of gun possession and their norms around carrying guns.”

Crucially, while shootings spiked in cities where crack was introduced, the violence was not limited to conflicts over drug turf. The three economists looked at murders of family members and intimate partners by young black men, along with suicides among the same population during the years after crack arrived. They found sharp increases in both fatal shootings of loved ones and gun suicides — but no similar increase in suicides or domestic murders by other means. “The increase in gun-related domestic violence murders shows that the increased availability of guns changed the technology of settling disputes and hence increased the murder rate,” Moore and his colleagues wrote.

In recent years, a fresh uptick in homicides has raised alarms again. The national murder rate rose by 10 percent in 2015, then another 8 percent in 2016. This July, the Centers for Disease Control and Prevention issued a research brief highlighting how firearm homicides fueled the rise, jumping 31 percent during those two years. While the federal government hasn’t released its 2017 figures yet, data from private sources suggests fatal shootings edged upward again last year.

Could a new influx of guns also explain the latest spike?

Some experts think so. According to the ATF, domestic gun production reached an all-time peak of 11 million weapons in 2016, fueled by the easing of local carry laws and inflated fears of new federal gun control. The single largest share of those weapons were semiautomatic handguns. The pistols flooding the market today are typically of higher quality, fire more powerful rounds, and have ammunition magazines with greater capacity than those made by the Ring of Fire companies.

At the same time, the recent homicide increase has been most acute in some of the same cities that had been ravaged by the introduction of crack three decades earlier, including Chicago, Baltimore and St. Louis.

“There’s nothing controversial about saying that means influence injuries,” said Dr. Sandro Galea of the Boston University School of Public Health. “Take the likelihood of people committing suicide by jumping off bridges. Studies show that when you make it harder to jump off a bridge, fewer people commit suicide that way. The supply of guns is an integral part of that same story: the widespread availability of a means of injury results in a greater amount of that kind of injury.”

As Galea sees it, the big difference between the rise in gun crime during the 1980s and ‘90s and the spikes in gun production and homicide we’ve just seen is that the earlier wave was followed by a dramatic federal policy response.

There is no such equivalent effort today.

“We are introducing more lethal means,” Galea said, “without any effort to mitigate their consequences.”