Roughly one in four dollars of nonbank financial intermediation in the U.S. today is conducted by an entity whose parent owns a regulated bank, a share that has grown steadily over the years, according to a new staff report from the Federal Reserve Bank of New York published July 15, 2026. The research, titled "Regulatory Arbitrage Within the Firm," documents how a substantial portion of what policymakers typically consider "nonbank finance" actually takes place inside bank holding companies, conducted by nonbank subsidiaries that operate alongside regulated commercial banks under common ownership and integrated management.
Nonbank subsidiaries account for roughly 20 to 30 percent of consolidated bank holding company assets in the post-2010 period, with broker-dealers making up the largest single component. While the largest dollar volumes are concentrated in the biggest and most complex bank holding companies, roughly two-thirds of bank holding companies with between $1 billion and $10 billion in consolidated assets operate at least one nonbank subsidiary. The spike in nonbank activity around the financial crisis reflects the conversion of Goldman Sachs and Morgan Stanley to bank holding company status in 2008, but even absent that episode, the broader trend shows nonbank activity has become increasingly embedded within banking organizations. Seventy-three percent of nonbank subsidiary-quarters in the sample are wholly owned by the parent bank holding company, and an additional 17 percent are majority-owned, so that 90 percent of nonbank subsidiaries fall under direct parental control. This ownership structure gives the parent effective authority over subsidiary capital allocation, dividend policy, and financing decisions.
The report finds that prudential banking regulation is centered on capital requirements that apply to the depository institution and to the consolidated holding company, but not to nonbank subsidiaries individually, creating what the authors call regulatory asymmetry. The median bank subsidiary holds equity equal to roughly 10 percent of assets, while the median nonbank subsidiary has an equity-to-asset ratio of 69 percent. The authors write that the median nonbank subsidiary accounts for a little under 2.5 percent of consolidated assets but nearly 16 percent of consolidated equity, for a multiplier of 6.5. According to the authors, nonbank affiliates function as "equity reservoirs: small balance sheets holding a disproportionate share of the organization's redeployable capital."
The research shows how this capital structure enabled a sharp divergence after Basel III's minimum capital requirements became binding for U.S. banks in 2015. Before 2015, excess capital—capital above all binding regulatory minimums—evolved in parallel for banks inside bank holding companies with nonbank subsidiaries and banks inside bank holding companies without them. Beginning in 2015, the two series diverged sharply: banks affiliated with nonbank subsidiaries accumulated substantially larger capital buffers, with the gap widening to 8 to 10 percentage points by 2020. At the consolidated holding company level, by contrast, no comparable divergence appeared; excess capital for bank holding companies both with and without nonbank subsidiaries declined gradually as the new requirements phased in. The report traces this pattern to equity reallocation within the firm itself, showing that parent equity investments in banks average 10 to 11 percent of bank assets over 2005 to 2024, while non-equity investments such as loans and receivables total just 1 to 2 percent.
The authors say they will use historical variation in interstate banking deregulation to identify differences in organizational complexity in the next post of their three-part series, tracing how holding companies reallocated capital across subsidiaries after Basel III. Their evidence points to much of the increase in bank capital coming from the equity reservoirs inside the organization and not from outside investors. Banks became better capitalized, but the organizations that owned them did not—a pattern the authors describe as regulatory arbitrage within the firm.