Chapter 20 Exercises

Work these with the chapter's central instinct: in small commercial, the unit of work is the book and the rule set, not the individual file, and the discipline is knowing which risks the machine should write and which it must refuse to write alone. Items marked with a dagger (†) have worked solutions in Appendix: Answers to Selected Exercises; the rest are for discussion or self-test. Section references like (§20.4) point you back to the relevant part of the chapter.

A. Recall and definitions

  1. † Define the business owners policy (BOP) in one sentence, and name the three core coverages it bundles. (§20.1, §20.2)
  2. Distinguish a BOP from a package policy (CPP) on three axes (who it is built for, how it is coverage-structured, and how it is underwritten). (§20.2)
  3. † Define straight-through processing (STP). What single phrase in the definition distinguishes it from a model that merely advises an underwriter? (§20.4)
  4. Define small-commercial class underwriting. Why does it price the class rather than the individual? (§20.3)
  5. State what automated/algorithmic underwriting actually relocates rather than eliminates, and where it relocates it to. (§20.4)
  6. List four coverages that a standard BOP generally does not include, naming the chapter that owns each. (§20.2)
  7. Name the three layers of the STP pipeline between intake and bind, and say in a phrase what each does. (§20.4)

B. The segment, the product, and eligibility

  1. † Explain the "upside-down pyramid": why small commercial is the broad base of the commercial market by one measure and the narrow tip by another, and what that single fact forces about how it is underwritten. (§20.1)
  2. A carrier's BOP eligibility guideline lists "auto repair — eligible up to 5,000 sq ft, no body/paint operations." Explain how this one line is itself an act of underwriting, and who it is being applied to. (§20.2)
  3. † For each of the following, decide whether it is a likely BOP risk or a package-policy risk, and give the deciding reason: (a) a 1,800 sq ft gift shop in a strip mall; (b) a 60,000 sq ft plastics manufacturer; (c) a two-room accounting office; (d) a contractor with heavy completed-operations exposure; (e) a small café with limited cooking. (§20.2)
  4. The chapter says the BOP often includes business income automatically, partly solving the Chapter 19 "limit nobody calculated" trap. Explain how that standardization both helps the typical small business and under-serves a business with an unusual income exposure. (§20.2; Ch. 19)
  5. Why is classification accuracy described as a "first-order control" in small commercial, when it is just one input among many on a large account? (§20.2, §20.3)

C. Underwrite this submission (route and decide)

  1. † Underwrite this submission. A submission arrives through the portal: "Marketing consultant," eligible class, good territory, no prior claims, model score clean and ready to auto-bind. But the address resolves to a light-industrial park, the insured reports \$400,000 of business personal property and a loading dock, and requests a property limit far above a typical consultant. State what the referral logic should do, which signal should catch it, and what you suspect is really going on. (§20.5)
  2. A small restaurant applies for a BOP coded "limited cooking." What two or three facts would you want confirmed before letting it auto-bind, and which knockout or grey-band trigger protects you if the coding is wrong? (§20.2, §20.5)
  3. † A 4,000 sq ft, four-person machine shop near Harbor Steel — manual lathes and a drill press, no welding, \$600,000 of equipment, clean five-year history, good protection class, not in the coastal cat zone — comes in through the portal. Walk it through the four-step STP pipeline and state, at each step, why it should ultimately auto-bind. (§20.4; The Underwriting File)
  4. A coastal small retailer, eligible class and clean history, scores well — but its address is inside the carrier's distance-to-water auto-bind cap. Should it bind, decline, or refer? Which kind of rule is doing the work, and why is that rule a hard knockout rather than a grey-band referral? (§20.5)
  5. Three submissions land at once: (a) an ineligible class; (b) a clean, standardized, in-appetite office; (c) an eligible class in a marginal score band with an unusual coverage request. Route each to decline/refer/auto-bind and name the signal family that governs it. (§20.5)

D. Price this risk / the economics

  1. † Price this risk (the economics). A BOP earns \$1,600 of annual premium. The fixed cost to underwrite, issue, and service it by hand is about \$300. (a) What share of premium is consumed by that fixed handling cost? (b) Compare to a \$250,000 account with the same \$300 cost. (c) Explain, in one sentence, why this single comparison is the economic case for automating small commercial. (§20.7)
  2. A small-commercial book runs a loss ratio of 62% and an expense ratio of 41%. (a) What is the combined ratio, and is the book profitable on underwriting? (b) Name the one lever most available to fix it in small commercial, and why it is the expense side, not the loss side, that this chapter emphasizes. (§20.7; Ch. 3)
  3. † A class rate is 5 points inadequate and is bound automatically across 100,000 policies. Explain, in words (no precise dollar figure needed — keep it qualitative), why this is a far more dangerous error in small commercial than the same five-point inadequacy on a single large account. Name the theme it illustrates. (§20.7; Ch. 1, Ch. 11)
  4. Efficiency and profitability are not the same thing. Construct a short example of a small-commercial book that is highly efficient (low expense ratio) and still unprofitable, and say what went wrong. (§20.7)

E. Find the red flag

  1. † Find the red flag. An STP program's monthly report shows that one eligible class — "metal products, light" — has grown to 8% of the book and its loss ratio has climbed for three straight quarters while every other class is flat. Nothing in the rules has changed. Name at least two distinct things that could be happening, and state the disciplined response. (§20.5, §20.7)
  2. A competitor known for loose underwriting just tightened its filters. Over the next quarter your auto-bind volume in several classes jumps. Why might this be a red flag rather than a success, and what would you check? (§20.5; Ch. 1)
  3. An agent complains that your portal refers "too many" of their accounts to a human while a rival auto-binds them all instantly. Is the agent's complaint, by itself, evidence your referral logic is too tight? What would you look at before loosening it? (§20.5, §20.6)

F. Write the memo / recommendation

  1. † Write the memo. In 150–200 words, write the recommendation that routes Harbor Steel to a commercial package policy rather than a BOP. Address class, size, complexity, and catastrophe, and state the one-line routing principle a junior underwriter should take from it. (§20.2; The Underwriting File)
  2. Draft a short referral-trigger specification (5–7 bulleted rules) for a new BOP program covering small retail and offices: list the hard knockouts and the grey-band referrals you would build before going live, and one sentence on why you start conservative. (§20.5)
  3. Write a two-sentence note to a small-commercial agent explaining why one of their accounts could not auto-bind and was referred — phrased to preserve the relationship and the speed expectation. (§20.6)

G. Ethics and judgment

  1. † Ethics dilemma. Your data team shows that adding a particular variable to the BOP pricing model would improve the loss ratio measurably. On inspection, the variable is highly correlated with the racial composition of the insured's neighborhood, though it is not race itself and does carry some genuine risk signal. The premium pressure is real and the margins are thin. Walk through how you would think about whether to use it, what questions you would ask, and where the line runs. (§20.7; Ch. 4, Ch. 35)
  2. The commercial pressure for speed pushes constantly toward loosening the filters and auto-binding more. Describe the discipline that resists this without losing the speed that wins the market — and explain why "automate the easy risks so you can afford to be careful on the hard ones" is an ethical posture, not just a commercial one. (§20.6)
  3. An automated decline goes out on a small-commercial submission because a third-party report flagged the applicant. What does the carrier still owe the applicant, and why does "the algorithm did it" fail as a defense? (§20.4 Compliance Corner; Ch. 8)

H. The Underwriting File and synthesis

  1. † Underwriting-File extension. Take the small machine shop from the chapter's contrast and turn it into a one-paragraph "Read the Submission" block in the six-field format (THE SUBMISSION / THE CONTEXT / WHAT IT SHOWS / WHAT IT DOESN'T / THE DECISION / THE LESSON), ending with the routing lesson it teaches. Label it as a constructed teaching example. (§20.2; The Underwriting File; Ch. 19 format)
  2. List the specific reasons Harbor Steel fails the BOP eligibility test, grouped under class, size, complexity, and catastrophe. For each group, name the chapter where that piece of Harbor Steel's program is actually underwritten. (§20.2; The Underwriting File)
  3. The chapter claims STP can be "genuinely better than human judgment" for the standardized middle of small commercial — a stronger claim than the book usually makes about automation. Explain the precise conditions under which that claim holds, and the precise conditions under which it fails. Why is small commercial the line where automation comes closest to replacing the individual decision? (§20.3, §20.4, §20.5; Ch. 1)
  4. Synthesize Chapters 19 and 20 into a single routing rule: given any commercial submission, what sequence of questions decides whether it goes into the automated BOP lane or onto an underwriter's desk as a package? State the rule in five questions or fewer. (§20.2; Ch. 19)
  5. (Looking ahead.) Chapter 31 (data-driven underwriting) and Chapter 32 (predictive modeling) deepen the machinery this chapter introduced. Name one thing this chapter said about the limits of automated underwriting that you expect those chapters to revisit, and why the limit is about judgment rather than data. (§20.4, §20.5; preview of Ch. 31, Ch. 32)