Since the implementation of the Solvency II Directive in the European Union in 2016, the quantification of lapse risk has become fundamental requirement for life insurers operating under this regulatory regime. It is now recognised as one of the three most material risks faced by life insurance companies.

Interest in lapse risk extends beyond regulatory compliance. Industry practitioners, particularly actuaries, have devoted significant effort to identifying and quantifying its potential impact, as well as in designing strategies for monitoring and mitigating this risk. Policyholder behaviour—especially decisions to terminate contracts—affects nearly every aspect of insurance operations, including product design, underwriting, profit testing, and asset-liability management, among others. Accurate modelling of lapse and surrender behaviour requires a clear understanding of the factors driving such decisions. This requirement constitutes the main motivation for the present study.

Our review reveals several important limitations in the existing literature. In some influential studies —including (De Giovanni, 2010), (Russell et al., 2013), (Finke & Mulholland, 2014), (Giri, 2018), (Gemmo & Götz, 2016) y (Wei Zhenyu; Zheng Haitao; Zhang Wenrui, 2019) —the type of insurance product under analysis is not specified, which limits the generalisability of their findings.

In other studies, such as (Eling & Kochanski, 2012), (Kuo et al., 2003) and (Eling & Kiesenbauer, 2014), the concepts of lapse and surrender are used interchangeably, despite referring to distinct phenomena: lapse typically denotes contract termination (or cancellation) without surrender value (common in pure protection products), whereas surrender (also referred to as withdrawal) refers to termination accompanied by a surrender value, as observed in savings, mixed, or insurance-based investment products (IBIPs). Although the term "lapse" is often used generically to describe early termination, we argue that distinguishing between these two behaviours is essential when identifying their respective drivers. Throughout this paper, lapse will refer to termination without a surrender value, and surrender will denote termination with such a value.

Our analysis suggests that the underlying drivers of lapse and surrender differ significantly across product types.

In addition, few studies adequately consider contextual variables such as the existence and maturity of secondary insurance markets, applicable tax regimes, macroeconomic conditions, or the availability of public pension schemes. Therefore, we consider the classification of studies by country to be essential for the appropriate interpretation and comparison of their findings. As noted by (Dar & Dodds, 1989) drawing empirical conclusions without accounting for the specific product and country-specific contexts entails a risk of misinterpretation.

The aim of this article is to consolidate existing knowledge on policyholder lapse and surrender behaviour, providing a structured foundation for future research. To this end, we identify and discuss the most relevant studies in each category while avoiding comparisons across fundamentally different contexts. By adopting this approach, we aim to reduce the risk of misapplying prior findings and to support practitioners in developing more robust, context-specific predictive models.

The remainder of this paper is structured as follows: Section 3 describes the methodology employed for the bibliographic review and presents the classification framework applied to the reviewed literature. Section 4 discusses the main findings of the study. Section 5 outlines the main conclusions, limitations, and proposes potential directions for future research. In addition, this section provides a brief overview of the main actuarial applications of the study. An appendix summarises the key hypotheses identified across the literature.

The research, was conducted through a systematic literature review, aimed at addressing the following research question: What are the determinants and triggers of life insurance termination?

The search strategy was initially implemented using the Web of Science database, employing the following search terms: "Policyholder Behaviour", "Surrender" and "Lapse". From the results obtained, only documents written in English and specially focused on life insurance were considered, irrespective of the country of origin or publication date.

The review was further restricted to publications classified within the following subject categories: Business Finance, Mathematics and Interdisciplinary Applications, Statistics and Probability, Mathematics and Behavioural Sciences.

This initial analysis identified a corpus of 113 publications, including scientific articles, theses, and other academic works. To ensure broader coverage of the literature, an additional 70 documents were subsequently incorporated, identified through their recurrent citation in the previously selected studies despite not having emerged in the initial search results. Following a preliminary assessment of 183 documents, a classification was carried out according to the primary focus of each study, with the aim of excluding those whose focus did not address the research question.

The excluded documents mainly addressed the economic valuation of surrender options (69 documents), optimal surrender modelling approaches (8 documents), behavioural analyses (9 documents), and miscellaneous topics outside the scope of the present study (12 documents).

After the aforementioned classification process, the final sample comprised 85 documents focused on identifying the determinants of life insurance policy terminations. These documents, developed for different purposes and across diverse products and markets, constitute the empirical foundation of the present study.

Following, the final sample were classified according to the criteria outlined below:

The analysis is further complemented by a synthesis of the principal variables incorporated in the models examined in the literature.

In addition, special attention is given to studies that do not clearly distinguish between types of insurance products or between the concepts of surrender and lapse. This distinction is particularly relevant in research addressing multiple products or mixed contracts.

All the hypotheses referenced in the analysis of results are described in the appendix.

The earliest studies focused almost exclusively on mixed insurance products. Over time, however, the literature has progressively expanded to encompass a broader range of product categories. In recent years, IBIPs has emerged as the most frequently analysed product type within the academic literature.

Table 1. Temporal Evolution of Publications by Product Type / Authors’ Own Elaboration

Of the 85 documents analysed, 9 do not specify the product type and refer to life insurance in general. These include (De Giovanni, 2010), (Russell et al., 2013), (Finke & Mulholland, 2014) and the bibliographic research on life insurance terminations (Shamsuddin et al., 2022).

Similarly, 2 of the 15 studies analysing multiple product types, (Giri, 2018) and (Poufinas & Michaelide, 2018) do not differentiate their conclusions by product category.

Furthermore, within the 26 studies examining mixed insurance products or not explicitly identifying the product type (17 focusing on mixed products and 9 without product specification), 15 do not differentiate between lapse and surrender. In particular (Kiesenbauer, 2013) and (Eling & Kiesenbauer, 2014) citing (Kuo et al., 2003), note that the term "lapse" is used to refer to both surrender and lapse, considering it consistent with the standard measures of analysis of termination rates.

The most widely considered hypothesis is the IRH. It predominates primarily in studies on IBIPs and, to a lesser extent, in research on traditional savings and mixed products. It is generally not considered in studies focused exclusively on protection products, although it occasionally appears in analyses covering multiple product types.

The EFH is more prevalent in mixed insurance and in studies encompassing various product categories. This finding is consistent with its rationale, as the need for additional liquidity can explain both surrender decisions and lapses resulting from an inability to continue premium payments in protection policies.

The combination of both Traditional Hypotheses (IRH and EFH) predominates in studies analysing multiple product types, and is particularly frequent in products with surrender value.

Overall, 59 of the 85 studies incorporates at least one of the Traditional Hypotheses.

Table 2. Traditional and Non-Traditional Hypotheses by Product Type / Authors’ Own Elaboration

Non-Traditional Hypotheses are more commonly observed in protection products. Within this category, as well as in mixed insurance, the Health Condition Hypothesis (HLH), and the Secondary Market Hypothesis (SCH) are the most prominent, particularly in the countries where a secondary insurance market exists.

In traditional savings insurance, the most frequently used alternative approaches combine the Level of Rationality Hypothesis (LRH) , with the interest rate hypothesis (IRH) and Policy Data Hypotheses (PDH), especially those related to expenses and the presence or absence of profit sharing. The Seasonality Hypothesis (SEH) is also supported in three of the reviewed papers. IBIPs share a large proportion of the Non-Traditional Hypotheses applied in savings insurance, with particular emphasis on the policyholder Level of Rationality Hypothesis (LRH), which is modelled on the degree of risk aversion, although it does not account for other factors that may influence such aversion. The Tax Impact Hypothesis (TXH) has been identified in only three of the papers analysed.

Regarding the hypotheses in the 15 studies which analyses several product types, most adopt a combined framework incorporating both the IRH and the EFH. In products with surrender value, the IRH tends to predominate, whereas in protection products, the EFH is more prevalent.

Table 3. Main Hypotheses in Studies Covering Multiple Product Types / Authors’ Own Elaboration

The United States clearly dominates the existing body of research in this field. Specifically, 30 of the 85 articles analysed focus exclusively on the Unites States market, plus an additional 5 joint studies examine multiple markets that include the United States. Consequently, the characteristics of the United States market should be carefully considered when interpreting the literature in order to avoid potential misapplications of the findings.

Within Europe, Germany accounts for the highest number of publications, with 9 papers identified, in addition to one study conducted jointly with the United States. Germany is followed by Italy, the United Kingdom, France, and Switzerland in terms of research output. In the case of Spain, four papers were identified. Furthermore, no studies were found for 17 of the 27 European Union member states, including the Netherlands, Portugal, and Poland, among others. In Asia, China, India, and South Korea emerge as the leading contributors, with three papers each.

Table 4. Studies Classified by Product Type and Geographic Region / Authors’ Own Elaboration

(*)The number of studies reported for Asia and Australia corresponds exclusively to research conducted within a single country. In contrast, the figures for the United States and Canada, as well as for Europe, also include joint studies involving multiple countries within the same region. The remaining joint studies are based on intercontinental analyses.

Three papers reviewing the state of research on policyholder behaviour were identified. One of them presents a bibliometric analysis, while the other two focus on identifying the main drivers and hypotheses considered in the literature.

Table 5. Summary of Literature Review Articles on Life Insurance Termination / Authors’ Own Elaboration

Table 6: Main Variables Considered / Authors’ Own Elaboration

This article provides a comprehensive review of academic literature addressing policyholder behaviour in relation to lapses and surrenders in life insurance, covering all product types across 85 publications over recent decades. To the best of our knowledge, this represents the most extensive review conducted to date in terms of product types and the number of studies examined, excluding purely bibliographic studies.

The study examines the approaches used in the literature and incorporates a substantially broader range of hypotheses than those considered in previous review studies, encompassing more than 20 key hypotheses.

In addition, the analysis differentiates the results by product type distinguishing between savings-and pure risk products, while also accounting for both geographic and temporal scope. This approach makes it possible to assess whether country-specific characteristics or particular economic conditions were considered when accepting or rejecting a given hypothesis. Furthermore, the review highlights the principal variables incorporated into the analysed models.

The main conclusions, areas for improvement, and potential areas for future research are presented below.

The results reveal a strong geographic concentration of studies in developed markets—particularly the United States, Germany, and Japan—while research focusing on Latin America and other emerging markets remains limited. This imbalance restricts the generalizability of findings and underlines the importance of expanding research to underrepresented regions.

Many papers provide insufficient contextualization regarding market conditions, legal environment, tax framework, or the existence of a secondary life insurance market, even though these factors that may have significant influence on surrender and lapse decisions. Furthermore, several studies do not clearly specify the type of insurance product under analysis or fail to disaggregate their conclusions by product type. Additionally, the concepts of lapse and surrender are often treated as interchangeable—a practice that may be acceptable when examining overall termination rates but problematic when identifying behavioural drivers, which can differ significantly between these two types of policy terminations.

These shortcomings have sometimes led to the misapplication of findings. It is therefore essential to consider that the conclusions of all these studies can only be extrapolated to similar situations; otherwise, they could mislead future researchers.

Regarding the hypotheses evaluated, the most widespread in general in all insurance types are the traditional hypotheses, namely the Interest Rate Hypothesis (IRH) and the Emergency Fund Hypothesis (EFH), which generally assume a rational policyholder seeking optimal termination timing. More recent studies, however, have challenged this assumption, introducing behavioural factors and alternative hypotheses. The inclusion of policy-level data and policyholder characteristics has proven particularly important, with many variables repeatedly identified as significant, although the set of significant variables varies according to the type of insurance product analysed.

Among the alternative hypotheses, the most frequently observed include the Inflation Impact Hypothesis (IH), mainly identified in mixed products; and the Tax Impact Hypothesis (TXH), primarily present in IBIPs products. Interestingly, neither of these hypotheses is commonly analysed in the context traditional savings products. Other relevant strands of literature include hypotheses related to the impact of the existence of the Secondary insurance Market (SCH), particularly in countries where this market is present, —most notably in the United States—, as well as the hypothesis of the impact of Health Conditions (HCH) which are mainly associated with risk and mixed products.

Several studies covering all product types also highlight the use of policy and policyholder data either as direct predictors of surrender behaviour or as variables supporting the EFH framework.

Finally, hypotheses that consider a policyholder behaviour that is not perfectly rational (LRH) become increasingly prevalent, especially in recent studies and mainly related to insurance-based investment products (IBIPs). In these approaches, deviations from full rationality are typically modelled though utility functions or different degrees of risk aversion. Other approaches related to the rationality of the policyholder behaviour incorporate constant probabilities, random components, by segmentation into behavioural sections/levels or scenario-based modelling. Additionally, this component has sometimes been linked to cognitive ability, financial knowledge (CAH), or financial advice (FAH). The theoretical proposal by (Lombardi et al., 2012) of a simulation model based on interactions between agents constitutes the most comprehensive conceptual contributions, however, due to its highly theoretical nature there is limited empirical evidence of its practical implementation, and the extensive data requirements likely hinder., its practical application.

In summary, an important direction for further research is to explore how individual characteristics and contextual interactions shape policyholder rationality, influencing their sensitivity to tax impacts, perceived product costs, investment preferences, or surrender timing. The extension of value-maximization models—such as the one by (De Giovanni, 2010) which incorporates an "irrationality" parameter based on economic indicators —like unemployment—could be enhanced using data on policyholders' financial distress, risk aversion, or investment behaviour. The practical implementation of models inspired by (Lombardi et al., 2012) also remains an open research frontier, with the potential to significantly improve surrender prediction in savings life insurance products.

The findings of this study have several important actuarial applications, particularly in improving the specification of policyholder behaviour assumptions used in the best estimate valuation of liabilities under both Solvency II and IFRS 17. The results may assist practitioners in identifying the key factors influencing specific insurance products, thereby enabling more targeted analyses and the adaptation of general assumptions to particular contexts, including a country’s economic environment, legal and tax frameworks, the type of surrender value(i.e. market value or actuarial reserve), and the distribution channel. This approach may help avoid excessive reliance on broad standard assumptions, such as the IRH alone. Such considerations are especially relevant in the assessment of extreme scenarios, which may pose significant risks to insurance companies. Indeed, the occurrence of such an event constituted the primary motivation for this research.

Furthermore, the results of this study may contribute to other actuarial areas by enhancing the development of more effective policyholder retention strategies, as well as strengthening Asset–Liability Management (ALM) frameworks through improved liability estimation.

MacKay, A., Augustyniak, M., Bernard, C., & Hardy, M. R. (2017). Risk Management of Policyholder Behavior in Equity-Linked Life Insurance. Journal of Risk and Insurance, 84(2), 661–690. https://doi.org/10.1111/jori.12094

These hypotheses are related to the economic value of the contract or to the policyholder's need for emergency funds. Their drivers are generally classified in the literature as either exogenous (e.g., unemployment rate, GDP, inflation) or endogenous (e.g., interest rates). Traditionally, the IRH and the EFH have been considered, although more recent studies have incorporated additional economic mechanisms into life insurance termination research. The main economic hypotheses considered in the literature are:

Derived from endogenous triggers:

Considered as rational, optimal behaviour or value maximization.

Proposed by (Dar & Dodds, 1989), is the most widely adopted hypothesis in the literature. It states that policyholders terminates their contracts in response to increases in market interest rates or stock yields, in order to take advantage of higher yields available in other investment opportunities. The hypothesis can be found in isolation within an framework optimal behaviour framework, for example in (Huang et al., 2022), (Gad et al., 2015) as well as in combination with other hypothesis— largely with EFH—, in studies such as (Eling & Kochanski, 2012), (Kuo et al., 2003) and (Cox & Lin, 2006).

In traditional savings products with guaranteed, fixed or variable rates, —particularly in Universal Life insurance—, the hypothesis is often expressed in terms of the spread between crediting rates and the market interest rates. In this context, it is referred to as the Crediting Rates Hypothesis (CRH), as discussed by (Vasudev et al., 2016).

This hypothesis considers the surrender as the exercise of a put option. It is generally applied to both variable and guaranteed annuities. It establishes the relationship between the surrender value and the value at maturity. Relevant contributions include (Escobar et al., 2016) (Kling et al., 2014)(Knoller et al., 2016)(Kent et al., 2009)(Alonso-García et al., 2024).

Based on empirical research by (Outreville, 1990) and (Russell et al., 2013), this hypothesis suggests that policyholders terminate an existing policy to replace it with another offering better prices or more favourable contractual conditions. Additional evidence can be found in (Bauer et al., 2017)(Fier & Liebenberg, 2013).

Derived from exogenous triggers:

Proposed by (Outreville, 1990), this hypothesis states that the policyholder decides to terminate the life insurance contracts when faced with adverse personal financial circumstances such as unemployment, health problems or long-term care needs. It appears frequently in the literature, both as a standalone explanation and in combination with other hypotheses. Representative studies include (Gemmo & Götz, 2016)(Liebenberg et al., 2012)(Belaygorod et al., 2014)(Dar & Dodds, 1989).

This hypothesis posits that the policyholders take into account the tax implications of the surrender amount when deciding whether to terminate a policy. It is not widely addressed in the literature. It has been identified primarily in studies on insurance-based investment products (IBIPs), although it may also be relevant to other product types depending on the contractual design and the tax regime of the country under consideration. Relevant contributions include (Moenig, 2021), (Moenig & Zhu, 2018), (Molent, 2019) or (Alonso-García et al., 2024).

Also referred to as the payment depreciation hypothesis, this approach is based on the concept of money illusion, according to which consumers may recognize the existence of inflation, but fail to perceive its real impact on their life insurance policies. For example, policyholders may interpret a constant premium as evidence that the policy maintains its value, even when the insured capital also remains unchanged in nominal terms.

An alternative interpretation establishes a positive relationship between inflation, household indebtedness and increased surrender activity. This perspective is based on the assumption that policyholders may require additional liquidity during periods of high inflation, thereby linking inflation to the EFH. Relevant contributions include (Wei Zhenyu; Zheng Haitao; Zhang Wenrui, 2019), (Cole & Fier, 2021) and (Babbel, 1979).

Proposed by (Cheng, 2022), this hypothesis combines two distinct effects depending on the type of insurance product considered: savings insurance with surrender value, and protection insurance. On the one hand, when a population-wide health crisis threatens, beyond death, unexpected medical expenses (particularly in countries without comprehensive public healthcare coverage), income reductions, and unemployment, produce an urgent need for liquidity. In this context, health shocks may contribute to higher surrender activity, acting as a trigger for the EFH.

On the other hand, a widespread health crisis increase policyholder’s awareness of the cover provided by their protection insurance. Consequently policyholders may become more inclined to maintain their contracts rather than terminate them.

This set of hypotheses often serves as a catalyst for economic hypotheses, and is therefore seldom analysed in isolation. For instance, an individual who is starting a family (e.g., profile of a young married person with small children) may require emergency liquidity, which can act as a trigger for Emergency Fund Hypothesis (EFH). Similarly, a policyholder with higher Cognitive Ability or greater Financial Knowledge (CAH) may behave more rationally and identify the optimal moment to surrender the policy, thereby, acting triggering the IRH or the MNH.

This hypothesis considers that certain policyholder characteristics—such as their age or marital status—as well as specific life events (e.g., retirement, divorce, transfer, death of a spouse, acquisition of a house, etc.) may influence the decision to terminate an insurance policy. For example, the reduced need for beneficiary provided by protection insurance, may lead to policy termination, an effect referred to as the Legacy Hypothesis by (Fang & Kung, 2012). These events may also act as triggers for additional liquidity needs under the EFH thereby, justifying surrender decisions in savings or mixed products, as discussed by (Liebenberg et al., 2012) and (Sirak et al., 2015).

In other studies, policyholder characteristics are analysed more generally in order to identify relationships between these attributes and the likelihood of early policy termination. Examples include (Mall & Sahoo, 2015) and (Milhaud et al., 2011).

This approach can be also found in combination with the policy data analysis, as illustrated by (Kueker et al., 2014) and(Yakob et al., 2018).

This hypothesis examines the impact that the policyholder’s level of rationality on the identification of the optimal moment to realize the profitability of the contract. It commonly associated with utility functions or degrees of risk aversion, as in the studies (Moore, 2009) and (Moenig, 2021). In other approaches, rationality is modelled as a constant component as in (Jia et al., 2023), (Kim, 2005a) or (Kuo et al., 2003), as a random component, as proposed by (Russo et al., 2017).Other studies classify policyholders into different sections or levels of rationality as in (Le Courtois & Nakagawa, 2013), (Li & Szimayer, 2013) and (Kling et al., 2014). Additional contributions analyse alternative scenarios as proposed by (Albizzali & Geman, 1994) and by the simulation agent based theoretical approach proposed by (Lombardi et al., 2012). The hypothesis is also related to Cognitive Ability/Financial Knowledge Hypothesis (CAH).

This hypothesis appears in studies related to savings or mixed insurance; however, it is particularly prevalent in the analysis of insurance-based investment products (IBIPs), where it is typically combined with economic hypotheses involving endogenous triggers.

This hypothesis associates a higher level of financial knowledge or greater cognitive ability with an increased probability of identifying optimal moment to terminate the contract. Cognitive ability is measured through survey-based methodologies in the study by (Nolte & Schneider, 2017). Similar hypotheses are also considered by (Giri, 2018) and (Pinquet et al., 2011).

A related proposition is the Financial Advice Hypothesis (FAH), also identified by (Nolte & Schneider, 2017) . This hypothesis states that, in situations involving liquidity needs, access to financial advice reduces the likelihood of policy termination. Furthermore, it suggests that financial advice is particularly relevant when policyholder’s level of financial education is low.

Originally proposed by (Dassios & Zhao, 2011) and later considered by (Barsotti et al., 2016), this hypothesis states that the policyholder behaviour is not independent, contrary to the assumptions adopted in most traditional life insurance termination prediction models.

According to (Barsotti et al., 2016) conventional modelling techniques allow the incorporation of multiple risk factors as covariates including tax relief, contract and policyholder characteristics, company reputation, regulatory frameworks, and financial market conditions. These approaches, therefore, generate heterogeneous predictions depending on the policyholder’s profile. However, from a global perspective, they fail to provide sufficiently accurate predictions in changing economic environments, as they assume independence in policyholder behaviours, and consequently, ignore potential contagion effects among policyholders.

Applicable to protection insurance and, due to their protection component, to mixed insurance products, this hypothesis states that some policyholders terminate their insurance policies because they perceive their level of risk to be lower than initially expected. Conversely policyholders tend to maintain their coverage after experiencing a health-related problem. Both effects imply that the remaining insured population presents a higher average level of risk than that considered in the initial underwriting assumptions. Consequently, the premium established for the original risk pool, may become insufficient for the redefined group, giving rise to what is known as reclassification risk. This hypothesis is analysed in several studies, including those by (Hendel & Lizzeri, 2003), (Finkelstein et al., 2005), (He, 2011), (Gottlieb & Smetters, 2014), (Pinquet et al., 2011) and (Bacinello et al., 2009).

This hypothesis establishes a relationship between specific policy characteristics and the decision to terminate the contract. It is identified in studies covering a wide range of insurance types. Frequently analysed variables include the duration, the policy age, profit sharing mechanisms (see (Eling & Kiesenbauer, 2012)), as well as the expenses and charges applied to the policy. In the latter case, it is referred to as the Charges and Expenses Hypothesis (CEH), as discussed by (MacKay et al., 2017) and (Bernard et al., 2014).

Other formulations of the hypotheses poses that high commissions, as well as long-term life protection policies and level-premium structures, —where the policyholder pays higher premiums at younger ages and lower premiums over time—reduce the probability of termination, as shown by (Hendel & Lizzeri, 2003).

Formulated in different variants, this hypothesis links policy termination to the financial or reputational situation of the insurer, establishing a positive correlation between increases in both magnitudes, essentially in crisis periods, as discussed by (Wei Zhenyu; Zheng Haitao; Zhang Wenrui, 2019)(Yu et al., 2019).

According to (Chang & Schmeiser, 2022) lapse risk increases when the insurer faces liquidity constraint. Moreover, extreme events may trigger a sudden increase in policy surrenders, which in turn can generate or exacerbate liquidity crisis for the insurer. As a result, both mechanisms may reinforce each other in a feedback loop.

This hypothesis is discussed in several studies, including , (Fang & Kung, 2012), (Giri, 2018), (Sakálová & Peller, 2015) and (Subashini & Velmurugan, 2015). It argues that higher termination rates are associated with policies sold where the policyholder with limited genuine interest in the contract, insufficient understanding of its characteristics or under pressure exerted by the salesperson.

In the literature reviewed, this hypothesis is more frequently associated with mixed insurance product, although, it may be considered applicable to a broad range of insurance product types.

This phenomenon is present only in countries where such markets exist, mainly in the United States, although it is increasingly observed in European countries such as Germany. The secondary insurance market typically operates through two main mechanisms: Life Settlement and Policy Loans.

A Life Settlement is a financial transaction in which the policyholder sells their mixed life insurance (typical in Whole Life insurance) to a third party for a one-time cash payment. This payment is lower than the policy's death benefit but generally higher its cash surrender value. The acquiring company assumes responsibility for future premiums and becomes the new beneficiary, collecting the insured capital if the insured person dies within the coverage period.

Policy Loans operate in permanent life insurance such as Whole Life or Universal Life insurance. In this mechanism the policyholder obtains a borrowed loan amount lower than the policy’s insured capital without undergoing a credit check or formal loan approval process. The loan is secured by the policy's cash value. If the borrower fails to repay the agreed amount, or in the event of the death of the insured person the lender becomes the beneficiary of the life insurance policy and would collect the cash value.

The Hypotheses related to the existence of a Secondary Market (SCH), as discussed by (Bauer et al., 2017), (Gatzert et al., 2009) and (Fang & Wu, 2020) suggest that the presence of a secondary market, is associated with a lower rate of termination rates. This effect is also identified in (Jiang, 2011) where it is referred to as the crowding out effect.

The Alternative Fund Hypothesis (AFH), closely related to the policy replacement hypothesis (PRH), argues that policyholders tend to rely on policy-backed loans on policies when alternative sources of credit are more expensive (Cole & Fier, 2021). Consequently, this mechanism is also associated with lower termination rates. Principio del formularioFinal del formulario

Several studies including, (Kim, 2005b), (Kim, 2005a), (Kim, 2010) and (Milhaud, 2013), suggest the existence of a seasonal patterns in life insurance terminations.

The literature in which this where this hypothesis is identified focuses exclusively on traditional savings insurance products