A regulation is assessed the way a project is assessed, by comparing what it costs with what it buys.
What goes into a regulatory cost-benefit analysis?
Regulatory cost-benefit analysis counts the compliance costs borne by firms, the administrative costs borne by the state, and the indirect costs borne by everyone through higher prices and foregone activity, then sets them against the monetised benefits. The hard part is never the arithmetic. Benefits are frequently non-market goods, meaning a life not lost, a river not poisoned, a crisis that did not happen, and those have to be valued by revealed-preference or stated-preference methods that are themselves contested. Costs, meanwhile, are concrete, immediate and loudly represented, while benefits are diffuse, delayed and represented by nobody in particular.
A cost-benefit analysis that finds a rule unjustified may be telling you about the rule. It may equally be telling you about the valuation method.
Net regulatory burden is the private compliance cost minus the private benefit, computed for the regulated firms alone. It is the number an industry association quotes when it says a rule costs the sector some figure. Treat it as a legitimate input rather than the answer, because the whole point of most regulation is that the private and social accounts differ.
Sunset clauses set an expiry date so a rule has to be re-justified rather than persisting by inertia, and regulatory impact assessments require the analysis before the rule rather than after. Both are procedural, both are widely mandated, and both are widely reduced to form-filling, which is a finding to check rather than assume.
Advertisement
How do you put a regulation into a valuation model?
Now the part that belongs to you rather than to the regulator. An analyst prices a regulation into a model in exactly one of two places, and the choice is not free.
Route one is the scenario: build the cash-flow effect explicitly for the years the rule is expected to bind, with a stated probability, and let the discount rate alone. Route two is the discount-rate adjustment: leave the cash flows alone and add a premium for the risk. Route one is right for a specific, dateable, largely idiosyncratic rule, which describes almost every regulation a single company faces.
Route two is right only for risk that is systematic, in the strict sense that it moves with the market and therefore cannot be diversified away.
Practice mostly uses route two because it is one keystroke, and practice is mostly wrong. The arithmetic in Worked Example 2 shows why. A tariff shock expressed as a scenario costs a US fastener maker 14.86% of its value; the same shock expressed as a permanent discount-rate premium requires 866.58 basis points, a number nobody would write down and which quietly assumes the tariff lasts forever. Quote the discount-rate equivalent only to make the size legible, and put the number in the cash flows.
{
"skill": "Building a merger screen and a regulatory-risk row from primary documents",
"use": "Have an assistant extract the share table, the notification thresholds and the remedy language out of a competition authority's published order or a company's own filing, then write the Python that computes concentration, the change in it, the critical loss and the diversion screen, so you can spend your time on market definition rather than on arithmetic.",
"tools": ["Chat assistants with code execution - Claude, ChatGPT, Gemini", "Python: pandas for the share table, plain arithmetic for the indices", "The authority's own order or decision PDF, and the primary statute or guideline text"],
"prompt": "Here is a table of firm names and market shares, a stated contribution margin, and a stated market-level demand elasticity. Write Python that computes: (a) the concentration index before and after a stated merger, (b) the change in it BOTH by subtraction and by the 2*s1*s2 shortcut, with an assertion that the two agree to 1e-9, (c) the four-firm concentration ratio before and after, (d) the critical loss for a 5% price rise and whether the candidate market survives the test at the stated elasticity, and (e) the diversion-based upward-pricing-pressure index for each merging firm under proportional diversion. Assert that the shares sum to 100. Do not round any intermediate value. Then state, in plain words, which single input the clearance decision is most sensitive to.",
"verify": "Recompute the index change by hand for the two merging firms before you accept anything: 2*s1*s2 takes ten seconds and catches a mis-keyed share table instantly. Trace every share to a named primary source (the authority's order, the company's own filing, or an industry association's published volume data) with its date, and state the denominator that share is a share OF, because a share of revenue and a share of volume are different numbers in the same market. Check the concentration thresholds against the CURRENT guideline text rather than a remembered band; those bands have been revised and a model trained on older text will state the old ones with complete confidence. Never let the tool choose the market definition: it will pick the one implied by the table you gave it, which is the assumption under examination.",
"diy": "You must be able to compute the index, the change, the critical loss and the diversion screen with a calculator on a five-firm table, and to say which way the verdict flips when the market widens. Worked Example 1 is the drill. A tool can produce all five numbers; it cannot tell you that site-mixed concrete belongs in the market, and that judgment is worth more than every number after it.",
"market": "IN"
}