The obligations I help an insurer see and fund

At the start of our review, a defined cohort of overdue reinsurance receivables totalled JPY4.6 billion. Repeated requests had produced activity without a sufficiently clear account of what would release payment. Claims colleagues were dealing with current policyholders, reinsurance specialists were managing contract relationships and finance was carrying an ageing number into the cash discussion. The executive team needed a view that joined those responsibilities without pretending they were interchangeable.

I sponsored a joint review of the material items, using the original cohort throughout so new receivables would not distort the result. We separated missing evidence from contractual disagreement and administrative delay. Each item had a responsible team, the next event required and an expected review date. Where a settlement or interpretation was needed, the relevant specialists retained authority. Finance could not declare an amount collectible simply because the forecast needed it.

The work cost JPY72 million over fourteen months, including temporary document support and specialist review. Some managers wanted to concentrate entirely on new business and current claims. I argued that collecting an existing asset deserved resources, while accepting that the work had to be prioritised by amount and practical progress. We did not instruct claims colleagues to delay policyholder payments while waiting for recoveries.

Collections from the opening cohort reached JPY3.3 billion, leaving JPY1.3 billion outstanding. There were no write-offs or offsets in that stated reduction; it was a cash-collection measure. The residual balance remained subject to its own accounting and collection review. I did not call the receipts new profit, because the receivable had already been recognised. Nor did the improvement establish that every future recovery would arrive on schedule.

The lesson was to record the reason for delay when an item first becomes overdue. We initially spent too much time reconstructing earlier correspondence. A better intake record would have made the next event clearer and reduced repeated requests that looked diligent but did not move the claim towards payment.

An actuarial review identified a JPY2.4 billion increase in the gross claims estimate for an older commercial-liability cohort. The movement concerned business written in earlier periods, not a sudden rise in current-year premium volume. In this bounded example, the company retained the full increase without an associated reinsurance recovery. Finance therefore showed a JPY2.4 billion reduction in pre-tax earnings before other movements, separately from the cash payments expected over time.

Underwriting colleagues wanted the annual performance presentation to emphasise stronger recent pricing. That improvement deserved attention, but it did not cancel the older obligation. I worked with the actuary and claims director on a bridge that showed the old-year movement and the current underwriting account separately. I asked what evidence had changed, which assumptions remained uncertain and what the next review would examine.

The chief executive challenged whether the explanation would make the overall business appear weaker than its current trading. I accepted the need for context and rejected a net narrative that made the reserve movement difficult to see. The board received both accounts, with the actuarial basis available for challenge. My responsibility was financial clarity; I did not select a claims estimate to produce a preferred earnings result.

We recognised the revised estimate and retained the affected cohort for subsequent development reviews. The immediate consequence was a less favourable reported result and more explanation to owners. The consequence was not an additional JPY2.4 billion cash payment on the review date. Claims colleagues continued to assess and settle individual files through their normal professional process.

I would improve the earlier management pack by keeping the history of the cohort's assumptions more accessible. We could explain the latest movement, but directors needed extra preparation to understand how the view had evolved over several years. A dated account of that evolution would make future challenge less dependent on the memory of people present at previous meetings.

A specialist commercial portfolio had written JPY6.2 billion of annual gross premium before we decided to stop renewing it. The expected economics of new commitments no longer met our appetite after claims and service costs were examined. Underwriting leadership recommended the change, and I prepared its financial consequences with the claims and operations teams. The decision concerned future business; it did not cancel the policies and claims already on our books.

An early cost proposal assumed that most support could be removed as premium volume fell. I asked for the remaining workload by activity: open claims, correspondence, records, complaints and reporting. The resulting plan retained eighteen experienced colleagues and a JPY240 million first-year operating budget. Some staff would support the runoff directly and others preserve access to systems and specialist knowledge needed by the team.

Commercial colleagues worried that retaining the cost would make the exit look indecisive. I argued that an unfunded service tail would make it incomplete. We agreed that the team and budget would be reviewed against work remaining, rather than a fixed percentage reduction tied to the disappearance of new premium. The board accepted the smaller future book and the continuing operating commitment together.

In the first year, the open-file count in the monitored cohort declined from 3,100 to 2,240. That count was not a measure of the value of remaining liabilities, and we did not use it to claim that the same proportion of financial risk had disappeared. The service team remained in place. New developments and complex files could continue to demand substantial effort even as simpler cases closed.

The decision gave up a visible volume contribution before all of its costs could fall. I would involve people planning even earlier in another exit, because an experienced colleague's uncertainty about their future can become a service risk long before their departure. The economics of runoff include the ability to keep capable people doing work whose commercial launch story has already ended.

The proposed claims-finance replacement had a JPY1.1 billion investment envelope. It promised better reconciliation and less manual handling across older and current records. The business case assumed a broad rollout, but the team had not yet demonstrated that historical claim references would migrate cleanly enough to support recovery accounting. A single approval would have committed the organisation before that dependency had been tested.

I supported an initial JPY380 million phase covering a bounded migration, reconciliation controls and parallel operation. Technology colleagues defined the implementation, while claims and finance specified the evidence needed to trust the resulting records. The next funding release required a review of migrated balances, exceptions, support effort and the revised rollout cost. It was not triggered simply by reaching a calendar date.

The programme sponsor objected that staging would delay the benefit and weaken negotiating certainty with the supplier. Both concerns were real. We accepted a four-month later rollout decision and JPY64 million of additional parallel-operation cost within the first-phase budget. In return, the board received evidence about the highest-risk dependency before considering the rest of the commitment. We did not describe the unreleased JPY720 million as money saved.

In the pilot, unresolved reference mismatches fell from 8.1% of migrated records in the first trial to 1.4% in the final reviewed trial. The remaining exceptions required manual handling, and the percentage concerned that pilot population rather than the entire insurer. The result supported a more informed next decision; it did not establish that full rollout would reproduce the same performance or that every proposed efficiency had been earned.

I would define the cost of running old and new systems together more fully at the start. The initial discussion treated parallel operation as a short technical phase, whereas it depended on experienced staff who were already carrying production responsibilities. Staged funding works only when each stage includes enough resources to produce credible evidence, rather than an artificially cheap demonstration.

Our cash and expense plan treated temporary claims support mainly as a response after a large increase in workload. Claims leadership wanted a pre-agreed capacity arrangement so trained help could be available sooner. The proposed annual retainer was JPY95 million, with additional charges if the capacity was called upon. Finance needed to compare that visible cost with the operational limits of relying entirely on recruitment and procurement after an event.

I asked the claims director to define the work that could be delegated, the training required and the limits of the provider's authority. The arrangement did not transfer our coverage decisions or responsibility to policyholders. We also reviewed access controls, supervision and how to bring the temporary team into the insurer's own processes. My role was to make the commitment and its possible call-out cost clear in the financial plan.

Some executives preferred to retain the budget for other work because the capacity might not be used. I supported the retainer but required an annual exercise and a review of whether the named resources remained available. The board approved a paid capacity option. The arrangement neither reduced the underlying claims nor established that every service delay could be prevented.

During the first year, a surge led us to use forty-two temporary colleagues for eight weeks, with JPY58 million of additional charges. The monitored queue of document-complete files returned from 1,460 to 510 over that period. The workload mix and incoming volume changed, so I did not attribute the entire reduction to the arrangement. Nor did a shorter queue establish that settlement decisions were correct; quality review remained separate.

The experience justified a more specific capacity discussion, rather than a claim of universal payback. I would place greater emphasis on supervisor availability in the next exercise. Additional processors can increase the burden on experienced reviewers, and a plan that budgets only for the people arriving can miss the colleagues needed to make their work useful.