At President Joe Biden's direction, the eviction moratorium from the Centers for Disease Control and Prevention has been extended to March 31. But tens of millions of people owing back rent may still be forced out of their homes even with the extension. After months of wrangling, Congress allocated $25 billion in rental assistance. But as city and state governments race to distribute those funds, they are discovering that data about evictions is so poor that we don't know who is losing their homes where, and how to focus aid and outreach.

The federal government collects data on evictions from public housing authorities. But it has little to no eviction information on the private rental market, where the vast majority of American renters live.

Recent research from New America, a Washington-based think tank, found that one-third of U.S. counties have no available annual eviction figures. Entire states, as well as vast swaths of others, lack eviction data crucial to answering basic questions: How many households are evicted each year? Why are those people being evicted? How much back rent is due in these cases? Which neighborhoods are hardest hit?

Absent reliable, consistent data, the Biden administration and state and local governments will be flying blind this year — with billions of dollars more in rent relief to distribute, but few ways of tracking and aiding communities most at risk of eviction and homelessness.

Extending the federal eviction moratorium is a temporary, incomplete fix: The hardest-hit families could be more than a year behind on rent by the time protections are lifted, potentially saddling millions of people with debts impossible to repay without targeted rental assistance. In December, analysts at Moody's, working with the thin public information that's available, estimated that almost 12 million renters would owe an average of $5,850 in back rent and utilities this January. There isn't a system that allows us to efficiently target this subsection of people. We already know that a lack of data contributed to widespread problems in disbursing CARES Act housing assistance funds.

The solution? In the coming months, the government needs to create a federal eviction database to help track and address housing insecurity.

Such a database could monitor eviction filings and judgments in near-real time. It could provide granular analytics allowing municipal leaders, housing advocates, service providers and researchers to see where evictions are concentrated, how eviction rates change over time, which landlords turn to eviction most readily and how much is owed by tenants facing eviction. It should promote the privacy of tenants who have been evicted, by restricting access to identifying information.

The good news: Courts already collect this data as part of the eviction process. The bad news: They vary wildly in whether they digitize eviction data, enter it in a database and make that database publicly available. A federal eviction database could fix this.

Right now, the data trail created by evictions is often harvested by private firms that then use it to provide screening services to landlords who routinely shut out people who were previously evicted. Instead, the government could use that trail to identify those most in need. The Eviction Lab, led by the Princeton University sociologist Matthew Desmond, built a public national database that helped us understand the prevalence of eviction across the country from 2000 to 2016. It allowed the public to see the eviction crisis that existed well before the COVID-19 pandemic: 3.7 million eviction filings in a typical year. (Compare that with 2.8 million mortgage foreclosures initiated at the height of the Great Recession.) And it has allowed us to confirm that women and Black and Hispanic renters disproportionately face eviction — a pattern that has continued during the pandemic.

More recently, the Eviction Lab launched a tool for monitoring eviction filings in real time across five states and 19 additional cities. And the Eviction Tracking System has allowed us to assess the efficacy of eviction moratoriums.

But these data sets are incomplete, and exist only because of the voluntary efforts of activists, lawyers, reporters, court officials and data scientists from around the country. Even the most dedicated teams of housing policy experts are no replacement for the financial resources and legal power of the federal government.

The Eviction Tracking System, for instance, covers only 1 in 5 renter households. That's because not all counties make data available, and a nongovernmental organization cannot force them to do so. In short, it's nobody's job to provide this crucial housing information, so it doesn't get done.

It would be wrong to reduce the collection of this data to an intellectual exercise for housing wonks. Systematically mapping evictions empowers working people and the public servants on their side, who are in desperate need of stronger evidence to fight for them.

A federal database would, for instance, make it much easier to track whether eviction enforcement is in line with renters' legal protections. Thousands of evictions that are illegal, or at least legally suspect, are being carried out across the country despite the federal eviction moratorium, but we don't have a way to track or prevent them.

The Biden administration has a number of ways to establish a national eviction database. Congress could pass a law directing the Department of Housing and Urban Development to create one — the bipartisan Eviction Crisis Act, introduced in the Senate in December 2019, proposes just that. Development of an eviction database has support in the House of Representatives as well. Biden could also direct the department to establish a database via executive order.

Still, to have teeth, these regulatory measures would need funding. An eviction database requires a sustained investment in data collection, standardization and management. The previous Congress authorized initial funding for a feasibility study, but a long-term commitment is required.

Regardless of the path, the result of these reforms would be a database that is publicly accessible — so that citizens and elected officials can better understand housing insecurity in their communities — but is heavily anonymized so that evicted tenants aren't discriminated against by landlords or employers.

Passage of the Eviction Crisis Act in the coming months would be a clear first step. There will be more to do once we better understand the scope of this housing crisis. But we cannot fix problems that we cannot see.

Peter Hepburn is an assistant professor of sociology at Rutgers University-Newark; he developed and runs the Eviction Tracking System at the Eviction Lab, a source for real-time eviction filing statistics. Yuliya Panfil is the director of the future of land and housing program at New America. They wrote this article for the New York Times.