As presaged earlier this week by the top House Democrat with authority over big-tech antitrust policy, Sen. Warren has now released what both of them call a “Glass-Steagall act” to break up giant tech platforms such as Amazon, Google, and Facebook.  Sen. Warren has frequently introduced a “21st-Century Glass-Steagall Act” on grounds that Glass-Steagall had worked well for decades before “repeal” in 1999 sowed the financial crisis’ seeds.  Like it, a Glass-Steagall construct would, advocates argue, segregate big-tech data from big-tech commercial activities.  Would this solve the numerous safety-and-soundness and equality risks I’ve highlighted in recent reports?  I doubt it.  To see why, it’s worth understanding what Glass-Steagall really does, what happened to it, and how a far less permeable barrier is necessary if we want to make big-tech finance safe, sound, and equitable.

The Glass-Steagall Act’s potency has taken on mythic proportions since progressives opposed the 1999 Gramm-Leach-Bliley Act. GLBA does not, however, repeal Glass-Steagall.  The Depression-era law prohibited a single company from conducting both commercial and investment banking.  This prohibition remains in law, if not in reality, to this day.  What GLBA did was to recognize market reality and then govern it by allowing a parent holding company to own subsidiaries that separately engage in banking and securities activities.

The real battle in 1999 was not over this structural recognition of a reality in place since at least 1986, when Bank of America bought one of the nation’s largest securities firms under the full force of the Glass-Steagall Act as the Fed and courts then understood it.  Securities firms owning insured depositories also date back not only to grandfathered companies under Glass-Steagall, but also to the non-bank bank charters Congress was unable completely to eliminate in a 1987 law.  Even from the 1980s to 1999,  there was so much integration under Glass-Steagall’s nose that the real issue wasn’t what a bank versus an investment bank could do, but how much securities and asset-management business a bank could do within its bank charter.  The biggest battle in the 1999 GLBA debate thus wasn’t actually over the subsidiarization construct, but a fight between the SEC and Fed over who would regulate all the in-bank securities activities GLBA left untouched.  The SEC tried its hardest to “push out” all the legacy securities activities authorized by the banking agencies, but broke its lance on the question.  As a result, Glass-Steagall remains in full force within financial holding companies, but the differences between commercial and investment banking remain not only porous, but also testament to the power of regulatory arbitrage.

What does this mean for big-tech finance?  It depends on whether the legislation segregates activities as  some propose or if, as Sen. Warren plans, the new approach forces divestiture  of “commercial” activities so that the platform companies become giant data and cloud-service utilities without ancillary – and very, very profitable – business ventures riding atop the platform.  The Glass-Steagall moniker might apply to a ring-fencing or targeted divestiture effort, but the far bigger one Sen. Warren has in mind is about the overall construct of giant tech platforms, not the complex conflicts and risks embodied by the Glass-Steagall Act’s approach to big finance.

Could Warren’s plan work?  In terms of financial services, this seems unlikely unless the legislation is very, very careful with the complexities of what’s data and what’s a financial service.  Current law is already a minefield, making it hard to see why a big-tech “Glass-Steagall” would walk through it any better than bank regulators did as big banks broke down the barriers between commercial and investment banking so effectively as virtually to eliminate them.

Take, for example, the question of payment services. Facebook is reportedly toying with a cryptocurrency scheme to handle payments in India as the first step to a global remittance service.  From the U.S. perspective, this might make Facebook a money transmitter subject to a raft of rules (e.g., AML, consumer protection).  But, is a cryptocurrency money?  Payors and payees at each end of the payment stream get fiat currency, but the middle is clearly a very gray zone with very large safety-and-soundness and money-laundering pitfalls.

Are P2P companies “lenders” or “finders” – i.e., do they intermediate funds in ways that trigger borrower, anti-discrimination, privacy, and other safeguards or are they just providing a data-infrastructure service that absolves the company of an array of legal responsibilities?  Is robo investment advice securities advice or just a similarly operational channel that permits an intermediary to take a fee without any responsibility for risks in the business that generates it?  Who then is liable if retirement funds are gambled on an adviser’s highest-fee offering?  What would happen if platform companies gain the entry they want to the payments system without the bother of liquidity regulations?

So far, the fintech and big-tech finance framework is at best a patchwork.  This applies not only to the fundamental challenge of defining what’s commerce versus finance, but also to an array of consumer protections based on what type of company does what with a consumer’s money under which legal framework.  A recent Economic Equality blog post laid out all these questions.  Suffice it to say here that they’re hard ones to answer under current law.

Assume, though, that someone somehow defines finance, establishes a like-kind regulatory regime, and institutes all the safeguards deemed necessary for a big-tech platform company.  What would then become of regulated financial institutions?  Many are spending billions to redesign offerings to capture data and exploit new delivery models.  Would banks also have to ring-fence data from the uses to which they are put despite all the rules that apply to their charters?  Maybe, but the real world of politics almost never requires any company to stop doing what it’s doing on the day Congress decides no one should do it anymore.

Absent very careful drafting, a tech Glass-Steagall Act would leave Amazon plenty of room to redesign itself as a non-bank bank and permit JPMorgan to keep on keeping on with its giant tech overhaul.  Just as lawyers built fortunes busting Glass-Steagall, so too will palaces be built on the fees earned from arbitraging the living day lights out of a new-style law. The fact is that it’s very, very hard to retrofit barriers into pre-existing business models.

It’s also hard to define core consumer-finance services and then demand the same level of protection in them regardless of the offering firm, but at least it’s a ground-up modernization of consumer finance in the digital age with a prayer of success because it’s a lot harder to arbitrage.  The famous definition of pornography – you know it when you see it – applies to consumer finance:  if a consumer is at risk because money is taken from, lent to, or handled for the consumer, then the service is financial and protections should apply. A better Depression-era example than Glass-Steagall is the Federal Deposit Insurance Corporation Act which, in concert with state laws, make it difficult to take a consumer’s money in ways that mislead the depositor into high-risk investments. A fintech recently arbitraged this prohibition, but got caught red-handed because this transgression – inadvertent or not – is a lot easier to spot as current law is clear about what’s a deposit regardless of who tries to take it.