It’s hard to find anyone who isn’t for regulatory cost-benefit analysis (CBA). Indeed, Chair Yellen appeared to endorse it at least in principle when she agreed earlier this week with the Trump Executive Order. However, the Trump plan is striking in espousing what may well be a substantive approach to CBA instead of the formulaic analyses the FRB to date has deployed to defend its rules from CBA advocates. The President’s Executive Order eschews the platitudes of cost-benefit analysis instead to call for “efficient, effective, and appropriately tailored” rules. This is the right way to go about CBA, but it’s also a lot harder to do than simply toting up the easiest costs one can find against the most unqualified benefits that could be attributed to a rule and allowing a regulatory edict to fly through a cost-benefit analysis without a scratch. Specifics from the Trump regulatory review about meaningful CBA would make a major difference in the post-crisis framework without materially undermining it. Here’s how to make it happen.

In 2012, FedFin issued a report laying out the challenges we saw then to simple CBA. Then as now, many in the industry were pushing for CBA not only because they hoped it would make the burden of new rules less crushing, but also because it gives opponents a large hook from which to suspend rules they want dead — ask the SEC about its proxy-access rule. However, what one means by CBA determines how much it makes a difference.

Much to the annoyance of CBA advocates, the FRB is not under the SEC’s express CBA requirements. It thus has a lot more latitude to count costs and benefits as it likes. Numerous rules – check out for example the GSIB surcharge and TLAC – bear evidence to the fact that the FRB’s approach to CBA makes everything it does seem all the better. Other bank regulators approach CBA in a still less sophisticated way. For example, when the OCC comes before OMB to argue that its rules do not have significant economic cost, it often counts as costs only all the new forms a rule requires or systems it mandates, not looking past policies and procedures to assess whether any bystanders such as borrowers might get hurt by the strategic adjustments a bank makes to handle the real costs of a new capital or prudential requirement. There aren’t any costs, so goes this type of CBA, because banks adjust to them and thus the bank bears no cost to speak of.

To counter these never-lose CBAs, the Hensarling CHOICE Act includes a very different approach to CBA. It requires a massive, double-diligent calculation of every possible cost at every stage of implementation for every affected institution, market, and economy. This as intended would stifle any and all new financial regulations.

Is this what every banker burdened by all the post-crisis rules should fervently desire? I don’t think so. All rules are not the devil’s handiwork. It’s of course tempting to put a pitchfork on the post-crisis rules, but over-burdensome CBA without meaningful results will not lead to the efficient, effective, and tailored framework rightly sought by the Executive Order. New rules that could be caught in the CBA quagmire could otherwise replace older ones with clearer standards that retain prudential rigor but add more certainty and clarity. Others are needed to replace anachronistic standards with innovative rules that reflect technological and market change. Sure, one could simply eviscerate the entire rulebook and let the market rip, but I’m not sure bankers would love so wild a market nor, for that matter, might the rest of us.

So, how to do meaningful CBA? First, it has to be marginal – that is, one has to look not at each new rule as a thing in and of itself, but rather as a thing intimately intertwined with many other like-kind things. How many capital rules do banks really have to have? How well do all the new resolution plans ensure meaningful market discipline that then warrants regulatory relief? Does stress testing counterparty exposures mean there is little marginal value to single-counterparty credit limits? And, how many other rules could be made more efficient and effective as President Trump demands if they are better tailored not just to the size of the company subjected to them, but also to the purpose for which they are intended?

Marginal CBA would make a big different within the body of rules, but meaningful CBA must also recognize that costs can’t be borne by one part of a market without undue benefit to competitors unless these costs are evenly distributed among like-kind companies. Many new rules nominally subjected to CBA conclude that all of their costs are worth it because the rule means less chance of a systemic crisis. And, if I stay in bed, I’ve less chance of getting hit by a bus. However, I’ve also a lot more of a chance of getting bedsores, fat, and very bored by watching too much TV. Making banks very, very safe keeps them from running amok, but it clears a path for competitors who, absent like-kind rules, almost surely will. So much for less systemic risk.

Finally, meaningful CBA must assure that any rule that passes the first CBA rounds described above can also be effectively enforced first by internal management and, should that fail, by regulators or the courts. All too many of the rules I read focus far more on mandating one after another policy and procedure each and every level of a company has to put in the great big loose-leaf binder plopped every quarter before an increasingly exhausted board of directors. Even worse, many new rules focus to an astonishing extent only on policies and procedures and/or complex, often-untested models. Do banks and their boards know what’s expected of them in substance, not just form? How about examiners? If compliance means filling out a lot of forms or back-testing a bunch of models, not achieving defined prudential goals that can be validated over time through objective criteria, do we know how safe a bank really is even if it’s in full compliance with everything each of its regulators throws at it?

In short, CBA is no guide to quality regulation unless it tells us whether one more rules make a meaningful prudential difference, if a new rule on one part of the market has consequences elsewhere that frustrate the rule’s objectives, and if the rule has measurable mandates that mean more than yet another big file of still more filled-out forms. If that’s what Mr. Trump has in mind, bring it on.