三篇保險公司和再保險公司長期關係的研究

Abstract

在實務與文獻上，概念上皆認為分保公司與再保險人維持愈久的交易關係，有助於提升再保險契約或市場交易的效率。然而，關於再保險交易關係如何影響分保公司的實證研究，非常地稀少。至今僅有幾篇學術研究以理論方式推論或者利用實際資料檢定當再保險交易關係期間愈長時，是否對於分保公司的財務績效、再保險使用程度與破產風險造成影響。因此，當再保險交易關係期間愈長，是否會對分保公司的其他面向營運績效造成影響，是目前學術上尚未探討並且是值得探討的重要問題。\n不對稱學習(Asymmetric learning)是再保險交易關係對分保公司造成影響的主要機制，即再保險契約訂立之初，再保險契約交易雙方存在資訊不對稱，此時再保險公司並未清楚分保業務真實的風險程度，因此無法訂立公允的再保險價格。隨著再保險交易關係期間愈來愈久，再保險公司持續地獲得與累積其承保業務的理賠資訊，愈來愈清楚業務的風險性質，並將其資訊用於風險分類與再保險定價。因此，當再保險交易關係期間短時，再保險資訊不對稱程度愈高；然而，當交易關係期間長時，資訊不對稱程度愈低。換句話說，在不同再保險交易關係期間下，分保公司與再保險公司同時面臨不同程度的資訊不對稱現象，因此分保公司面臨不同的誘因並進一步做出對本身有利的決策，進而造成不同面向的影響效果。\n除此之外，文獻上捕捉再保險交易關係的指標也稀少，目前文獻發展的指標僅捕捉時間面向的關係，然而，在衡量再保險交易關係時，也必須將再保險交易的金額多寡納入考量，以避免在計算關係指標時納入太多交易金額太少的再保險交易，進而無法捕捉重要的再保險交易關係。因此，本文藉由考量過去五年累積分保保費金額的方式進一步捕捉各種重要性程度的再保險交易關係，即創造同時捕捉時間與交易金額面向的衡量再保險交易關係的指標。\n本博士論文由三篇探討再保險交易關係如何影響美國產險業分保公司不同面向的績效所組成，概念上，本論文想要回答以下三個問題，即分保公司的核保利潤、業務多角化與市佔率是否會隨著再保險交易關係期間的增長而隨之變動。換句話說，探討承保活動、風險管理策略與策略績效是否與再保險交易關係有相關。\n第一篇文章探討當再保險交易關係期間愈長時，分保公司的損失率、綜合率與核保利潤如何受到影響。藉由非線性變數的配置，一階段系統和差分動差法的迴歸模型估計，實證結果顯示，再保險交易關係和損失率與綜合率皆呈現倒U型關係，再保險交易關係和核保利潤呈現U型的關係。再保險交易關係短的時候，損失率和綜合率會隨著再保險交易關係的延長而增加，核保利潤則會隨之降低。當再保險交易關係久時，損失率和綜合率皆會隨著再保險交易關係的延長而降低，核保利潤則會隨之增加。因此，本文提出實證證據支持不對稱學習假說。\n第二篇文章檢視再保險交易關係如何影響分保公司的業務多角化程度，並且探討何種因素調節再保險交易關係與多角化程度的相關性。利用一般最小平方法與隨機效果Tobit模型估計，實證結果顯示，當再保險交易的關係持續時，分保公司的商品多角化程度會愈高，也發現公司規模將減緩上述再保險交易關係與多角化程度相關性。因此，本文提出實證證據支持不對稱學習假說與真實服務假說。\n第三篇文章則額外考量到市場上同時競爭的保險公司，探討再保險交易關係如何影響分保公司的市占率。藉由非線性變數的配置，一階段系統和差分動差法模型的估計，實證結果顯示，再保險交易關係與市占率呈現倒U型關係。因此，本文提供支持不對稱學習假說的證據。\n因此，本博士論文提供較完整的證據，回答文獻和實務上尚未回答的問題，即維持愈長期的再保險交易關係，會影響到分保公司哪些面向？再保險交易關係造成的影響效果，是否會隨著累積的交易關係年數而改變？本論文的實證結果可呼應再保險市場依賴長期交易關係的觀點，並且可進一步給實務人士、學術研究者、監理官和要保人等相關的利害關係人提供參考，以供進一步的決策依據。

It is well known in practice and in academic research that the efficiency of reinsurance transaction is on the long-term transaction relationship basis. However, how reinsurance relationship has an influence on cedants is an underexplored issue in academic research. Specifically, only few studies derive, theoretically, and examine, empirically, the effects of the tenure of insurer-reinsurer relationship on financial performance, reinsurance usage, and bankruptcy costs of cedants. The questions left to be answered in current reinsurance relationship literature are what other dimensions of cedants could be influenced by the reinsurance relation’s duration.\nThe underlying mechanism of reinsurance relationship is learning over time, called asymmetric learning phenomenon. Specifically, reinsurers could gradually learn the risk of ceded business over time. In addition, the measure of capturing the reinsurance relationship is scarce in recent reinsurance relationship studies. Thus, we also improve such measures by taking the amount of reinsurance premium ceded in the past 5 years into account to capture various level of importance of reinsurance relationship. That is, we construct the reinsurance relationship measure simultaneously considering time and transaction amount dimensions.\nTherefore, in this thesis, we present three essays on the effects of the tenure of insurer-reinsure relationship in US property-casualty insurance industry. Conceptually, this thesis intends to answer the questions of how underwriting profitability (underwriting activities), product diversification (risk management strategies), and market share (strategic performance) are affected with the increase of the tenure of insurer-reinsurer relationship.\nThe first essay investigates how the tenure of insurer-reinsurer relationship affects underwriting profitability, including loss ratio and combined ratio, of cedants. The results document that the tenure of insurer-reinsurer relationship is inverted U-shaped related with loss ratio and combined ratio, respectively. In addition, a U-shaped relationship between the tenure of insurer-reinsurer relationship and underwriting profitability is found. The results are consistent with asymmetric learning hypothesis.\nThe second essay empirically scrutinizes the effect of reinsurance relationship’s tenure on product diversification. The results show that cedants tend to diversify more with the increase of the relation’s duration. In addition, we also find that such effect is mitigated as the firm size increases. The results are consistent with asymmetric learning hypothesis and real service hypothesis.\nThe third essay, considering other rivals in insurance markets, empirically investigates the effect of the tenure of insurer-reinsurer relationship on market share. The result documents that such tenure is inverted U-shaped related with market share. The results are also consistent with asymmetric learning hypothesis.\nOverall, the implication of this PhD dissertation is that maintaining longer relationship with incumbent reinsurers is beneficial for cedants and various stakeholders, especially for those with long reinsurance relationship.