Bar insurance in Oklahoma

Oklahoma is the worked example on this site of detail not being evidence. A great deal is published about this state’s service rules and much of it cannot be stood behind. This page states what is verifiable, marks what is not, and declines to fill the gap with confident prose. Declining to state something is a finding in itself, and it is more useful to an operator than fluent guidance that cannot be checked.

Every other state page in this series opens on how alcohol service is licensed there and what the state does about liability arising from it. This one does not, and the reason is not that Oklahoma is a quiet state on the subject. One other page does not open that way either, for reasons of its own, and that is our Virginia bar page.

It is that the account assembled for this build could not be attested. The route by which it was obtained was never recorded, so nobody can now say what was read or when. And the authority its central position depends on — the thing that would settle it — was never actually obtained at all.

So none of it appears below. Not the licensing body, not the license types, not the liability position, and not a softened version of any of them. What follows is what is missing, why a confident summary is precisely the wrong thing to trust, and what an Oklahoma operator can usefully do about it this week.

Detail is not evidence, and this page is the worked example

It would be easy to assume that a page withholding a state’s liquor position is doing so because the material behind it was vague, or hedged, or obviously shaky. That is not what happened here, and the fact that it is not is the most useful thing on this page.

The account assembled for Oklahoma is specific. It is confidently worded. Set beside a properly researched state page it would sit there quite comfortably, and nobody could sort the two apart by reading them. That is the problem, stated as plainly as it can be put.

Two things are wrong with it and neither is visible from the text. The route by which it was obtained was never recorded, which means nobody can now say what document was read, or when, or whether it was current at the time. And the authority its central position actually depends on was never obtained — the thing that would settle the question was not, in fact, consulted.

Put together, those two defects do not produce a slightly less reliable account. They produce something that is not an account at all, dressed in the clothes of one. Fluency is free to produce; it costs nothing to write a paragraph that sounds settled, and nothing in the paragraph itself reveals whether anybody checked. Precision is not a proxy for accuracy, and the more precise an unchecked account is, the more completely it will be believed.

So the four subjects that would carry this page are absent from it: which body licenses alcohol service in Oklahoma, what kinds of license exist and what each permits, what position the state takes on liability arising from service, and where that position comes from. Nothing above is asserted and nothing is denied, and nothing should be read out of the order the four are listed in. If you finish this page holding a belief about any of them, it did not come from here.

What to require of the next account you read

The generalisable version of this is a habit worth having, because you are going to read other accounts of this subject and most of them will read better than this page does.

Ask what the account rests on, and require an answer that names a thing rather than a category. “The statute” is not an answer; a section number is. “Case law” is not an answer; a decision is. An account whose author can point at the specific text they read is in a different position from one whose author is summarizing a summary, and the difference is invisible unless somebody asks. It is worth seeing the test met: our Ohio bar page turns on a numbered section of that state’s revised code and names it in every paragraph that leans on it.

Then ask when it was checked, and against what. A position that was accurate at some point and has not been re-examined is not a wrong answer — it is an unknown one, and unknown is the honest label for it.

And notice what the answers to those two questions do to the tone. Somebody who has genuinely read the source will usually tell you where it was awkward, what it did not cover, and where they would want a lawyer. Somebody working from a confident secondary account has nothing awkward to report, because summaries smooth the awkward parts out first. Unbroken confidence about a state’s service-liability position is a reason to ask another question, not a reason to relax. The opposite habit is visible elsewhere in this series: our Tennessee bar page sets down two statements from two sources, declines to say what they add up to, and points at the sections instead.

Applied here, it means putting the licensing question to the body that issues the credential rather than to any account of it, describing your room as it actually trades rather than as it reads best, and putting the liability question to an Oklahoma lawyer. Then hold both replies to the two tests above before you build anything on either.

The version of this page we did not publish

Withholding is easy to describe and hard to feel, so be concrete about the alternative. The other version of this page exists. It opens by naming the licensing body, sets out the license types and what each permits, states the state’s position on liability arising from service in a paragraph that reads as settled, and turns to the room. It is a good page, and nobody finishing it would think to ask a question.

Follow what it does after you close it. You believe it, reasonably, because it is specific and well written. You stop looking. You describe your room to a broker in the vocabulary it handed you, make a limit decision against the position it stated, and perhaps train the floor to the rule it described. Every one of those rests on an account whose route nobody recorded and whose central authority nobody obtained — and none arrives carrying a caveat, because a caveat is read once and the substance is acted on for years.

The cost is not that the account might be wrong; it might be entirely right. The cost is that you would have no way of telling and would have stopped asking. This version costs you something too: work the other page would have appeared to do for you. Put the licensing question to the body that issues the credential, and the liability question to an Oklahoma lawyer.

What it does not cost you is a file. Everything you know about your own room is untouched by the outstanding question: what hours are kept, when the room fills and how long the peak runs, what the staffing looks like at that moment, whether entertainment is booked and how often, and whether the same faces come back or a new set arrives every week.

Nor is the rest of the program affected. General liability, property and workers compensation are ordinary lines on ordinary terms here, and the Oklahoma property and payroll conversation is on our Oklahoma restaurant page. What a liquor liability policy is and how it sits against the general liability form is on the liquor liability page, and how a bar is underwritten as a trade rather than a jurisdiction is on the bar insurance page. Every one of those rests on material somebody actually retrieved.

And one sentence to your broker closes the rest of the gap: the state position has not been established for you, and you want the program built without an assumption standing in for it. It takes thirty seconds. The page we did not publish would have made it feel unnecessary.

Major Oklahoma bar markets

• Oklahoma City. The larger of the state’s two metropolitan trading areas, and the place where the missing licensing vocabulary costs a submission most, because it is where the widest range of operating formats sit next to each other. Describe the room by its hours and its crowd rather than by a category name this page cannot verify for you.
• Tulsa. The state’s other metropolitan center, and a separate trade area rather than a satellite of the first. An operator with rooms in both should have the licensing question answered for each of them rather than once, because nothing on this page establishes whether the answer is the same in both.
• Norman. A market whose weeks are shaped by a calendar rather than by an ordinary weekly rhythm, which means the busiest nights are predictable and the quiet ones are genuinely quiet. Age discipline is the one control worth running to a fixed standard here regardless of how the outstanding legal question is eventually answered.
• Stillwater. A smaller trading area where demand concentrates on a handful of dates rather than spreading through the year, so the operation running on the busiest night is not the operation described by an average week. Describe the peak explicitly — it is the part of the file that no licensing vocabulary would have carried anyway.