Over the past two decades, audit regulation in the United States has expanded dramatically, becoming more detailed, more centralized, and more prescriptive than ever before. In the wake of major accounting scandals, like Enron in the early 2000s, lawmakers responded with sweeping reforms, most notably the Sarbanes–Oxley Act and the creation of the Public Company Accounting Oversight Board. These changes fundamentally reshaped how audits are conducted, inspected, and enforced. Yet, despite their scale and cost, a basic question remains surprisingly hard to answer: Which audit regulations actually work, at what cost, and why?
In a new white paper, Clive Lennox, professor of accounting at USC Leventhal School of Accounting, argues that the core weakness of modern audit regulation lies not in its motivation, but in how it is designed and implemented. Major regulatory interventions are typically introduced across the entire audit market all at once, without prior testing and with little ability to evaluate their real effects afterward. Once a rule is imposed on every company and every auditor, there is no meaningful comparison group, making it difficult to determine whether observed changes are caused by the regulation itself or by unrelated economic forces. As a result, regulators, lawmakers, and the public are often left debating regulatory effectiveness without clear causal evidence.
The paper proposes a shift in mindset: new audit regulations should be treated less as a set of fixed commands and more as a series of testable policy experiments. Instead of immediately imposing new rules on all public company audits, regulators should, whenever feasible, introduce them through randomized pilot programs. Under this approach, a new rule would initially apply only to a randomly selected subset of public company audits, while others would continue under the status quo. This design, which is standard in medicine and occasionally used in other areas of financial regulation, would allow policymakers to observe what changes when a rule is introduced while controlling for outside economic factors.
Randomized pilots offer several practical advantages. First, they generate clearer and more credible evidence about whether a regulation improves audit quality, raises costs, alters behavior, or produces unintended consequences. Second, they reduce the risk of locking in ineffective or counterproductive rules at scale. Third, they strengthen regulatory legitimacy by grounding policy decisions in transparent evidence rather than assumptions — making those decisions easier to defend to courts, legislators, and the public. Importantly, pilot testing does not imply deregulation or inaction. It simply means learning about the effectiveness of proposed policies before extending them to the full population of public company audits.
The paper points to Regulation SHO, which governs short selling in equity markets, as a rare but successful example of this approach. In that case, regulators used randomized exemptions to study market effects before making permanent changes. Audit regulation, Lennox argues, has largely missed similar opportunities, despite affecting trillions of dollars in capital markets and imposing substantial compliance costs.
