The following post was submitted by Francesca Grifo, director of the Union of Concerned Scientists' Scientific Integrity Program, and Donald Mays, senior director for product safety planning for the Consumers Union.

The Consumer Product Safety Commission's (CPSC) recall of millions of toys unsettled a lot of American families last summer. Acutely aware of risk of making bad toy choices, many parents are now wondering whether the agency created by Congress to protect them from unsafe products is itself defective.

The easy explanation is that the CPSC is starved for resources and staff. If we only throw more money at it, things will be fine. That's partly true, and Congress is on the verge of pumping up the agency's budget. But more money and resources alone won't solve the real problem.

That problem is as hidden as the lead in Thomas the Tank Engine's paint or the magnets in Polly Pocket figures -- both which children could ingest. It's the culture of secrecy that permeates the agency, undermining its mission.

That secrecy prevents CPSC scientists -- the chemists, hazard statisticians, engineers and human factors experts -- from sharing their research with the public, or even with colleagues working outside the agency. And it compromises the agency's duty to protect the public from defective products. Equally troubling, it shields both the agency and the companies it regulates from accountability.

Federal law aids and abets this penchant for nondisclosure. A provision in the Consumer Product Safety Act requires the CPSC to obtain the permission of a product's manufacturer before it publicly releases any information about that product, or risk being sued. With the threat of a lawsuit hanging over its head, the agency often negotiates the wording of a recall announcement as painstakingly and slowly as two countries negotiating a peace treaty. Every word, every nuance is parsed.

Those who urge Congress to eliminate this debilitating provision face a Catch 22 situation. Secrecy about how the agency reaches decisions about product recalls, its negotiations with companies, and the timeline it follows makes it nearly impossible to prove that the provision actually suppresses important safety information or delays it from reaching the public. But it is apparent that the agency has used the law to cover up its own failures and censor its own scientists.

For example, in May 2005, a preschool teacher in Indiana called the CPSC hotline, alerting the agency to a horrible accident involving a 5-year-old in her class. He swallowed some powerful magnets that had fallen out of a Magnetix building toy. He needed emergency bowel surgery to save his life. She urged the agency to immediately recall the toy.

The CPSC failed to act. A few months later, on Thanksgiving Day, a toddler in Seattle died after swallowing magnets from another Magnetix toy. The agency did not recall Magnetix toys until March 2006.

A powerful series in the Chicago Tribune this year pinpointed these and other failures at the agency. When Tribune reporters asked agency officials why it had taken so long for the CPSC to act in a number of cases, the agency stonewalled them. Why? "In refusing to answer questions about Magnetix," the Tribune reported, "the agency cited a provision of federal law that protects manufacturers' reputations."

The fiasco over lead in children's lunchboxes also is emblematic. In September 2005, the CPSC stated that it had found no hazardous lead levels in 60 soft vinyl lunchboxes. When questioned about its testing methods, the agency refused to give out any information, citing its need to keep the names of the products' manufacturers confidential. It took a year, and a Freedom of Information Act request, to shake any information loose about the tests. The Associated Press and the Center for Environmental Health discovered that the CPSC tested fewer than 10 lunchboxes, and when it found high lead readings, it changed the way it tested until it got the safety results it wanted.

Protecting manufacturer confidentiality also gives the CPSC the leverage to muzzle its own scientists, forcing them to jump through bureaucratic hoops to publish any research, even research that doesn't cite any company names.

Likewise, it helps justify the agency practice of requiring departing employees to sign non-disclosure agreements, discouraging them from blowing the whistle on agency incompetence after they have left. It contributes to an environment where agency scientists often are not free to share their research with their colleagues in academia.

An agency that relies on secrecy cannot be held accountable. Nor can the companies it is supposed to regulate. It's a major defect in product safety law that Congress can and should eliminate.