【当該年度の実績概要】交付申請書に記載の研究目的および研究実施計画に沿って、２０２３年度に以下のことを実施した。 １．２０２２年度に調査取得した、個人の金融商品購買選好に関する調査データ（有効回答１０００件）を精査し、三種のリスク区分ごとにリスク・プレミアムの算定を行った。これらに基づき、各区分のリスク選好についての仮説（純粋リスクには強い回避選好、狭義の投機的リスクには穏やかな回避選好、射幸リスクには強い愛好選好を想定）を、期待値原理に基づくリスク定量化によって検証し、これらの妥当性を確認した。 ２．同じデータに基づき、純粋リスクにおけるリスク回避選考の強さと、投機的リスクにおけるリスク愛好選好の強さの相関を分析した。これらのリスクを定量化して回帰分析を行ったところ、得られた相関はごく小さかった。これらは負の相関があると考えるべき理由があるように思えたため、これに反する結果が得られたことの意味について考察を行っているが、先行研究が少ないこともあり、年度内には理由を理論づけるに至っていない。 ３．上記２．の結果が真のリスク性向を表すものであれば、興味深い結果と考えられる。これを確認するため、調査票設計上の問題に起因するバイアスでないかを精査しており、第二回、第三回調査の実施が遅延している。 ４．この状況を踏まえ、研究期間を一年延長し、第二回以降の調査の実施及び成果の論文化は、２０２４年度に延期する。（補助事業期間延長申請済み）
【２０２４年度の計画】１．調査票のバイアス検証および第二回・第三回調査を実施する。 ２．3区分ごとの市場のリスク・プレミアムの推定結果を学会誌に公表する。 ３．純粋リスクにおけるリスク回避選考と投機的リスクにおけるリスク愛好選好の相関（無相関に近いと想定）について、その計量結果と原因の分析を実施する。

保険加入等のリスク回避的選好は，効用関数が凹関数であることを示唆する。一方，宝くじや馬券の購入等のリスク愛好的な行動は，効用関数が凸関数であることを示唆する。両者が同時にみられることは，一様に凹または一様に凸である関数による効用の理解を困難にしている。この困難を解決するため，リスクを純粋リスク，投機的リスク，射幸的（宝くじ等）リスクに3区分し，それぞれの区分に異なるリスク選好が働くと仮定する枠組みが考えられる。この3区分の有効性を確認する目的で，小規模の予備的調査を行い，各選好を表すリスク・プレミアム（当該リスクを回避または追求（獲得）するために人が支払ってよいと考える費用）を測定した。その結果，純粋リスクを回避するためのリスク・プレミアムは大きく，投機的リスクでは相対的には小さいこと，また射幸的リスクでは，逆にリスクを追求（獲得）するためのリスク・プレミアムが大きいことが示唆された。小規模かつ統制の不十分な調査ではあるが，この結果は，上記3区分が有効である可能性を示唆し，さらに，本格的な検討の基礎となりうるものではないかと期待する。

This study introduces a novel framework that consistently explains people's risk preferences. The author aims to evaluate these preferences using this innovative framework quantitatively.The framework categorizes risk into three types: pure, speculative, and speculative (or lottery). This explains the seemingly contradictory mix of risk-averse preferences for insurance and risk-loving preferences for lotteries. This comprehensive approach will allow us to measure the risk premium for each of the three categories, providing novel insights into risk management strategies.This study categorizes risk into three types: pure risk (insurance), speculative risk (stock investments), and lottery risk (gambling) to capture varying risk preferences. Using survey data from 1,000 Japanese participants, the study assesses willingness to pay (WTP) for insurance, stock investments, and lotteries. The research introduces two metrics to quantify risk appetite: a risk premium coefficient (h) and risk premium per unit of standard deviation (RP/Δσ). The research plan involves conducting multiple surveys over three years. The survey was conducted in fiscal year 2022, yielding 1,000 valid responses. Significant findings include that individuals exhibit domain-specific risk preferences, with risk aversion predominant in insurance, risk-seeking behavior in lotteries, and moderate risk aversion in stock investments. Importantly, risk preferences across categories show minimal correlation. Risk attitudes are not universal; being enthusiastic about insurance does not imply that individuals are risk-averse regarding investments or gambling. Individuals demonstrate varying risk behaviors based on the nature of the risk. These findings suggest that economic models should consider context-dependent risk preferences rather than fixed personal ones. The results impact financial decision-making, insurance pricing, investment behavior, and the design of policies. As of the end of March 2025, the results and implications have been incorporated into a paper titled “Understanding Risk Preferences: A Three-Type Classification Approach.” This paper was accepted for the upcoming 2025 World Risk and Insurance Economics Congress in Calgary, August 3-7, an international conference co-hosted by the American Risk and Insurance Association (ARIA), the European Group of Risk and Insurance Economists (EGRIE), and the Asia-Pacific Risk and Insurance Association (APRIA). We aim to publish the results in The Journal of Risk and Insurance after getting feedback at the WRIEC 2025. Additionally, we believe these findings suggest further insights when the issue is explored in more detail, and we are currently developing the design of subsequent surveys to gain a broader understanding.

This study introduces a novel framework that consistently explains people's risk preferences. We aim to evaluate these preferences using this innovative framework quantitatively.The framework divides risk into three categories: pure, speculative, and shooting (lottery). This explains the seemingly contradictory mix of risk-averse preferences for insurance and risk-loving preferences for lotteries. This comprehensive approach will allow us to measure the risk premium for each of the three categories, providing valuable insights for risk management strategies.Our research plan entails conducting multiple surveys over a three-year period (initially over two years). The survey was conducted in FY2022, with 1,000 valid responses. As of the end of March 2024, we analyzed the survey data and measured the risk premium regarding expected value and variance pricing. We identified the distribution of subjects' risk premiums. We also conducted cross-sectional analyses and regression analyses between risk preferences for different types of risks. It showed that the magnitude of risk aversion in pure risk has almost no relation with the magnitude of risk love in shooting risk. We believe these findings are worth further exploration and are currently working on the design of subsequent surveys to further clarify the meanings of the findings.We aim to publish the results in foreign or domestic journals in FY 2024.

This study proposes a framework to explainpeople's risk preferences in a consistent manner and aims to quantitativelyevaluate risk preferences using this framework. The framework divides risk into threecategories: pure risk, speculative risk, and lottery risk. This explains theseemingly contradictory mix of risk-averse preferences for insurance andrisk-loving preferences for lotteries. The research plan is to conduct multiplesurveys over a two-year period and to measure the risk premium for each of thethree categories. This year's objectives were (1) to examinethe design of the survey, particularly the ease of understanding the survey questions by respondents,(2) to verify the validity of the method by conducting the first round of the survey,and (3) to obtain 1,000 valid responses.As of the end of March, (1) to (3) above progressedas expected. In FY2023, we plan to analyze the dataobtained from the first survey and measure the risk premium, design and conductthe second and subsequent surveys, analyze the data, and publish the results inforeign or domestic journals.