JP2004524594A - Method and apparatus for determining loan prepayment score - Google Patents

Abstract

No abstract.

Description

は申込者のデータおよびローンパラメータのベクトルである。
【数１０】

は、ロー・ディスクレパンシー・シーケンス（ＬＤＳ）をベースにした計量経済学的な一連のシナリオであり、それらはＲＴＨにリンクした指数計量経済学モデルによって生成されている。従って、そのモデルは、それぞれ繰上返済性向計算に関連する主要マクロ経済学指標の力学および相互作用について記述する一連の推計学的な微分方程式である。
【００３１】
分析的な繰上返済モデル
【外３】

（ローンのタイプに応じて変わる）は、申込者のデータ（Ａ）、ローンパラメータ（Ｌ）および計量経済学的なパラメータ（Ｅ）に基づいて所定のシナリオで繰上返済評価値ｐ_ｓを計算するように調整される。
【数１１】

【００３２】
シナリオｓにおいて時間Ｔまでに蓄積された繰上返済の合計は次のように計算することができる。
【数１２】

【００３３】
次いで、時間Ｔにおける繰上返済の合計は以下によって与えられる。
【数１３】

【００３４】
そして最終的に、繰上返済スコアは以下のようになる。
【数１４】

【００３５】
前記繰上返済スコアを生成する分析モデルは、後の研究にも基づく、追加の外部の行動に関するまたは計量経済学的ファクターによって、またモーゲッジブローカー挙動の前述した行動のスコアリングによっても、更に形成されても良い。
【００３６】
本発明は、クレジットエンジニアリングワークステーション（以下、ＣＥＷという）という形の代替実施形態にて表わされ得る。このＣＥＷ（以下により完全に説明する）は、ローンオリジネータが全ての繰上返済計算、モデル分析、および最初に述べた前記繰上返済モデルを用いる本発明の価格提示を行うことを可能にするユーザーインタフェースを備えている。
【００３７】
前記ＣＥＷは、オラクル、ＳＱＬサーバー、サイベース、ＤＢ２あるいはインフォミックスデータベースサポートを用いる、ユニックス（Ｕｎｉｘ（登録商標））またはウィンドウズ（登録商標）ＮＴ（Ｗｉｎｄｏｗｓ（登録商標）ＮＴ）の環境下で演算する。前記ＣＥＷは、ＪＡＶＡ（登録商標）／ＨＴＭＬブラウザーベースのグラフィカルユーザーインターフェースと共に、ＣＯＲＢＡ、あるいは構造化されたオブジェクトモデルをも用いる。
【００３８】
前記ＣＥＷのサブルーチンはすべて、消費者の繰上返済性向を決定する最終目標に寄与する。例えば、本発明取り引きのサブルーチンは、各種金利シナリオの生成、およびモデル化された金利シナリオを履歴上および現行利率利回り曲線パーフォーマンスおよび他のマクロの経済指標に適合させる後続の経済シナリオモデルフィッティング方法をサポートする。
【００３９】
このシステムの一部は、消費者繰上返済挙動に関する報酬の衝撃を効率的に測定し値付けするための報酬値付けロジックを含んでいる。例えば、それは、貸し主のローンを繰上返済しないことに対して消費者に報酬を与える貸し主に最も有益である。そのような報酬は、消費者繰上返済行為の影響に関して査定され得る。従って、このシステムは、末端ユーザーが形式上の報酬構成を設計すること、および予期される消費者繰上返済行為への影響をテストすることを可能にする。
【００４０】
各種のユーザー定義可能スクリーンも、デフォルトスプレッド、繰上返済スプレッド、ブローカーコミッションスケジュール、および消費者に提示されるべき商品の値付けに影響を及ぼす他の金融ファクターを確立する。各種の他の経済シナリオはユーザーインタフェースによって集められ、合理的に値付けされた末端ユーザーモーゲッジ契約における他の貸し主が定義した基準結果はもとより各種の見込みとデフォルトデータと共に組み合わされる。
【００４１】
図３には、本発明のＣＥＷに関する更なる情報が示されている。このシステムは、基本的なグラフィカルユーザーインターフェースであるユーザーインタフェースモジュール７０と、オリジネータが貸し主から金銭を借りたい消費者に関する情報を提供することを可能にする他のソフトウェアとを含む。このユーザーインタフェースモジュールは、ローン属性７６、申込者属性７４および報酬プログラム属性７２のコレクションを許容する。さらに、ユーザーインタフェースモジュール７０は、スプレッド、ブローカーコミッションおよびローン７８に関連する他のコストを収集しあるいは計算する。ローン属性７６および他のローン関連コストは、他の情報と共に、適切なローン価格８６を生成するのを支援するプライシングエンジン８４に供給される。
【００４２】
いずれもローンの価値に影響を及ぼすものであるローン属性７６、申込者属性７４および報酬プログラム属性７２は、繰上返済計算サーバー８０に供給される。繰上返済計算サーバー８０は各種繰上返済モデルパラメータからのインプットを受信し、繰上返済スコア８２を生成する。
【００４３】
図４は、繰上返済モデルを生成するのに必要な相互作用を示すブロック図である。申込者属性７４およびローン属性７６からなる消費者情報９６は、繰上返済モデルフィッティング９２モジュールに供給される。繰上返済モデルフィッティング９２は、繰上返済履歴データ９０に基づいた各種繰上返済モデルパラメータ９４を確立する。いったん適切な繰上返済モデルが繰上返済モデルフィッティング９２によって生成されると、モデルが、消費者へのローンの種類が要求する与えられた特定の消費者の繰上返済スコアの計算のために前記繰上返済計算サーバーに戻される。またこの繰上返済計算サーバーは、計量経済学モデルシナリオジェネレータからのインプットから利益を得る。
【００４４】
図５は、計量経済学モデルのための相互作用を示す。計量経済学モデルシナリオジェネレータ１０６は、計量経済学モデルフィッティングモジュール１０４およびＬＤＳシナリオ１０８からインプットを受信する。計量経済学モデルフィティングモジュール１０４は、計量経済学的履歴データ１００および現在のマーケット環境１０２から、金利の上昇や低下およびその傾向に関する情報（これらに限定されるものではない）を含む情報を得る。計量経済学的履歴データ１００からの情報は、消費者が所属する人口統計学的グループおよび年齢、収入、クレジット評価、職業や他の要素のような他の計量経済学的情報に関係がある。現在のマーケット環境１０２からの情報は、金利の変化の方向および速度に関係がある。計量経済モデルシナリオジェネレータ１０６はその情報を処理し、該情報に基づいて各種シナリオを生成する。
【００４５】
再び図３に戻って、繰上返済計算サーバー８０は、当該特定の消費者のための繰上返済スコア４４を生成する。繰上返済スコア４４は、確立された繰上返済モデルおよびその生成された計量経済学モデルに基づいている。繰上返済スコア４４はプライシングエンジン８２へ送信されて当該消費者へ提示されるローン商品の価格が決定される。
【００４６】
図６に、ユーザーインタフェースモジュールが各種シナリオを生成するために用いる追加パラメータを示す。本発明の追加の側面は、新商品の生成を提供する。ストラテジーオプティマイザー１２２は消費者による提示された商品の引受けに基づき、他の商品からのおよびそれに関するインプットがマーケットに出る。ストラテジーオプティマイザー１２２は、個人の貸し主のポートフォリオに基づいてマーケティング計画を生成する。そのようなマーケティングプランは、貸し主にとってより有益な市場への新商品の提供を支援することができた。このシステムは、最も好ましい繰上返済特性、つまり繰上返済の低い性向を有する消費者へローン商品を提供する途を提供するオプティマイザー１２４を含む。このシステムは、更に財政的に有利な消費者行動に報いる消費者への申し出および他の誘因をモデル化し定義するローヤリティオプティマイザー１２６をも含む。チャンネルオプティマイザー１２８は本発明の一部である。チャンネルオプティマイザー１２８は、各種金融商品を搬送する最も効率的な方法であるチャネルを評価し決定するために金融商品提供物の搬送チャネルを分析する。このシステムもまた、洗練された繰上返済履歴データ９０および計量経済学的履歴データ１００を継続的に作るために各種データベースの情報を受け取り、整理するデータベースオプティマイザー１３０を含む。
【００４７】
本発明のシステムが実行されるターゲットプラットフォームは、３２メガバイトのＲＡＭ、ハードディスク記憶装置および検索を備えたインテルペンティアム（登録商標）プロセッサーをベースにしたシステム、およびＴＣＰ／ＩＰプロトコルを用いたコミュニケーション能力のいずれかである。これに代えて、本システムは、サンソラリス（Ｓｕｎ Ｓｏｌａｒｉｓ）プラットフォーム上でＵＮＩＸ（登録商標）オペレーティングシステムの下でも実行する。いずれの場合においても、ユーザーのためのディスプレイは、ハードコピー報告書を出力する能力を有するものであることが期待される。典型的なオペレーションにおいては、このシステムサイトから離れた複数のユーザーは、プライベートネットワークあるいはインターネットを介してシステムにアクセスし、本発明が繰上返済スコアに結びつく希望の計算を行うのに必要な情報を送る。その後、このスコアは遠隔端末で、要求するユーザーに送られる。
【００４８】
繰上返済スコア生成のシステムおよび方法を説明した。いわゆる当業者は、本発明の他の改変は本発明の範囲を逸脱せずに可能であることを認識するであろう。
【図面の簡単な説明】
【図１】図１は本発明に係る方法の概要を示すものである。
【図２】図２は本発明のブロック図である。
【図３】図３はユーザーインタフェースモジュールコネクションを示すブロック図である。
【図４】図４は繰上返済履歴データを備えた相互作用を示すブロック図である。
【図５】図５は計量経済学モデルを備えた相互作用を示すブロック図である。
【図６】図６はユーザーインタフェースモジュールによって用いられるファクターを示すブロック図である。[0001]
(Field of the Invention)
The present invention relates to the acceptance and processing of applications for mortgage loans and other debt instruments. More specifically, the present invention allows investors to assess the value of the mortgage or other debt instrument, or to create a customized loan product tailored to the financial behavior of the consumer or borrower. To enable, an apparatus and method for assessing a borrower's prepayment propensity in the form of a prepayment "score" is provided.
[0002]
(Background of the Invention)
First, consider the most common type of bond, the consumer mortgage. Mortgage value mainly depends on the mortgage period. In the first phase of a mortgage, there are intermediary fees and various other settlement costs charged to the lender. If the mortgage is extended over time, the lender has the opportunity to recoup the cost of setting up the mortgage in place for a given consumer and benefit from that mortgage. This is particularly important for all businesses that lend money, but especially for mortgage lending institutions that have shareholders and other investors.
[0003]
When a mortgage loan is paid off earlier by refinancing, it may be possible for the lender to actually make a loss on that particular mortgage, depending on how soon the mortgage loan is paid off. The prepayment rate depends on many objectives. For example, during periods when mortgage rates are falling, more consumers generally refinance their mortgage loans to reduce monthly payments. However, given a given macroeconomic environment and other important and objective factors, each consumer will demonstrate an individual propensity to prepay the loan. This prepayment propensity reflects the demographic and other objective attributes of the consumer. A system that can assess such individual prepayment behavior by consumers prior to a loan can make a more profitable loan and therefore increase loan funding to more consumers or borrowers. Accordingly, the present invention applies without limitation to a) pricing mortgages and other debt securities, b) valuing existing portfolios of debt securities, and c) risk management of institutions that hold debt securities. obtain.
[0004]
Another element of the present invention is mortgage broker monitoring and scoring. Mortgage brokers deal with both consumers or borrowers and lenders or clients. It is possible for brokers to encourage their consumers or borrowers to refinance their mortgages frequently and prematurely to generate brokerage commissions. This causes the mortgage broker to generate its broker fee, but the early prepayment of earlier mortgage documents creates a loss for the lender. Thus, the present invention also has the ability to allow mortgage brokers to evaluate prepayment behavior.
[0005]
Broker behavior is not necessarily hateful. Consumers who adapt particularly to rising and falling interest rates may be those who change mortgage instruments more frequently than the average consumer. Brokers scored based on the consumer's prepayment behavior that the broker brings to the lender will want to know a consumer's prepayment propensity. This is useful because it allows a mortgage broker to optimize a broker's relationship with its lender or client by simply bringing in a consumer or borrower with a low prepayment ratio.
[0006]
Therefore, it is very important for lenders and brokers to be able to better determine prepayment behavior before marketing or underwriting fees are incurred. These payments are too large to be blindly absorbed on behalf of consumers with poor prepayment tendencies. Indeed, the invention can be beneficially used to manage the initial marketing effort itself. For example, only customers who can show a favorable score for prepayment behavior may receive an invitation for mortgage merchandise A. Consumers who have been identified as having inherent prepayment risk may be offered a more appropriate mortgage product B that reflects the increased risk. Thus, enhanced customer segmentation and product design initiatives lead to consumer interests and their debt financing.
[0007]
To understand the potential impact of the nationwide prepayment scoring standard, Fair, Isaac and Company, Inc., for over thirty years, as specified in the present invention, .) (Hereafter Fair Isaac) may refer to the existing default risk scoring criteria owned and distributed by Fair Isaac. By establishing and widely disseminating a standard methodology for scoring borrower default risk, Fair Isaac has dramatically increased the insight of mortgage lenders to the expected loan dynamics. In finance, enhanced insight is synonymous with enhanced information. Enhanced information means that the lender's risk is reduced. Finally, this reduced lender risk profile results in less capital loss. In other words, because Fair Isaac has successfully standardized an alternative measure of default risk, consumers can borrow more money at better and cheaper interest rates. Markets are more efficient than before, and everyone benefits.
[0008]
For further qualification of the timeliness of the invention, see Reference 1, Green Tree Chief Returns $ 23 Million ..., Wall Street Journal, March 1998. The story highlights industries that widely and precariously surround prepayment speeds in the consumer debt portfolio. Greentree Financial, a leading company in the industry, said, "We have suffered severely over the past year as bad debt has widened as a painful return of over-accounting. In addition, borrowers have been seeking lower interest rates, An unexpected wave of loan prepayments has hit the industry, indicating that working-class consumers were not as naive as lenders believed. " Even if the deducted profit is based in part on voluntary prepayment assumptions, Greentree will significantly exaggerate last year's earnings and exercise its option under GAAP accounting to reduce the estimated lending profit. Pushed ahead to gain. Green Tree miscalculated these prepayment speed assumptions, which forced the company to write off $ 390 million of its 1996 reported profit in 1997. In 1998, the company was sold to Conseco.
[0009]
Early disclosures in the area of prepayment scoring in a lending context are limited or fictitious. US Patent No. 5,696,907, entitled "Systems and Methods for Performing Risk and Credit Analysis of Financial Services Applications," was granted to Tom. The Tom patent discloses the use of a neural network that mimics the underwriting decisions of loan company executives. The Tom patent method provides a non-licensing standard that generates a useful criterion in preparing new or modified underwriting guidelines to increase profitability and minimize the loss of future portfolios of loans. Based on iterative regression method. The prepayment observation is used in the neural net as a negative flag, but the prepayment scoring system is not used in Tom's patent.
[0010]
In view of the prior art, there is a clear need to determine and predict a consumer's prepayment propensity and a clear and strong need for a method and apparatus for generating such measurement and predictive parameters.
[0011]
(Summary of the Invention)
The system and method according to the present invention generally functions as follows. The service bureau or broker receives individual loan applications from consumers electronically. These loan applications are sent to the lender for assessment. Lenders use the invention to submit their loan applications for review and analysis. The loan application is reviewed by the present invention according to a sophisticated economic and customer behavior model, and the prepayment behavior of the candidate borrower is scored. The scores of these borrowers, which are an index of their prepayment propensity, are returned electronically to the lender. The lender then uses the prepayment score to adjust the appropriate mortgage price, including interest rate, commission, broker commission and potential consumer compensation settings. Using this consumer scoring technique, lenders can strive to contact or contract with consumers who have a low propensity to prepay.
[0012]
Advanced scoring of customer prepayment probabilities greatly improves the lender's risk profile for newly lending customers. This new insight adds value to marketing, underwriting, financing and management methods for first and second mortgages, credit card carryforwards, and asset-backed loans such as car loans. By supporting the lender in customer segmentation with this critical behavior metric, wasteful and excessive costs are sent from the lending economics organization. In this way, more money is available at a lower price for more people.
[0013]
For borrowers, this system offers several advantages. First, more favorable loan terms are provided to consumers who exhibit beneficial borrowing behavior, i.e., borrowers who maintain the loan for a useful period without prepayment of the loan. In addition, dealing with a stable borrower market results in a more favorable financial environment for all lenders, thereby reducing the risk of deficit and broadening the financial benefits of borrowers in the normal course of all efficient markets. Is given.
[0014]
Again, the irrefutable economic relationship between financial risk exposure and expected financial rewards enlivens the environment addressed by the present invention. If lenders reduce their risk through increased prepayment scoring (and by their extension costs), the borrowing costs paid by consumers will ultimately fall.
[0015]
For a loan originator, this system offers several advantages. Loan originators can more efficiently price specific loans. In addition, the loan originator can more efficiently choose a broker or intermediary to select the best borrower. Furthermore, the systems and methods of the present invention provide more efficient direct and indirect sales investments by identifying individual and group consumers who exhibit the most beneficial borrowing behavior (ie, propensity to not pay off financial debt). Bring.
[0016]
Given the saturation of traditional direct channels (eg, email and outbound telephone transactions) and the surge in direct marketing costs, any available efficiency in the direct marketing method is highly desirable. For example, in marketing home equity lending limits (ie, second mortgages), direct mail response rates are currently on average less than 0.3% (ie, less than 3/10 of 1%). Obviously, even a small fraction of this small sample of respondents proves to be unsuitable for debt products being marketed in terms of prepayment behavior. Therefore, tailoring of certain debt instruments to consumers of certain prepayment behavior characteristics is necessary for efficient pricing of bonds. Lead generation, third-party data collection, underwriting, and yield variance calculations all form direct bond profitability and are all beneficially affected by the present invention.
[0017]
Finally, in the context of sophisticated asset and liability management (ALM), subtle prepayment behavior analysis offers significant benefits to its practitioners. As ALM seeks to minimize catastrophic asymmetries in asset and liability cash flows as a primary objective, intelligent risk managers will change the expected period to strengthen their balance sheet Use debt service contracts. For example, a lender's risk manager may assign multiple classes of bonds that reflect multiple prepayment profiles to ensure himself sufficient incoming cash flow to retain his expected liability cash outflow. You may look for it. Therefore, in reconciling the expected cash-in and cash-out flows, a thoughtful risk manager carefully evaluates the segmented portfolio of bonds, scored by prepayment propensity (and other measures) to prevent a liquidity crisis. Use.
[0018]
An additional equally beneficial use of the present invention is in assessing existing mortgage or bond blocks of a business. This assessment may be required by the lender risk manager, auditor, regulator or investor. It may actively reflect the interests of stakeholders in asset liability risk management or may be implemented as part of a merger and acquisition price assessment. In all instances, the prepayment scoring system quantifies the prepayment rate characteristics from granular perspectives to pools, ie, block perspectives, bonds. As seen in the Green Tree case, failure to properly assess prepayment risk has huge balance sheet implications and typically leads to overestimating the portfolio or the entity itself.
[0019]
For the auditor, the system of the present invention provides a quantitative measure of the prepayment risk and reduces the auditor's "collection" record. This situation occurs when the drawer that undertakes these mortgages accelerates earning based on a prepayment assumption that is under generally applied accounting procedures (GAAP). If their prepayment assumptions are incorrect, the last year's earnings report is incorrect and heavy charges are required to reflect lower portfolio earnings.
[0020]
To financial authorities, the system of the present invention offers the ability to quantify balance sheet risk due to expected consumer prepayment behavior. This allows regulators to more accurately measure and assign minimum bank capital levels.
[0021]
For credit rating agencies, the ability to score prepayment risk according to objective, standardized methodology provides significant assistance in assessing a lender's creditworthiness. Rating agencies function effectively as credit market leaders. Lenders rely on favorable credit valuations to lift their debt at favorable rates, and rating agencies carefully assess lenders' capital adequacy claims, as in the case of regulators. The capital held by the lender (cash reserve) is directly affected by the bond prepayment rate. This is because under GAAP accounting rules, lenders are allowed to obtain a substantial percentage of the expected future profits for a given shortened bond, and those gains themselves are: Because it essentially depends on the assumed life of the receivable (in the case of a subprime mortgage, for example, if the mortgage is kept valid for 4 years instead of 3 years, the profit may double) unknown). If those gains are exaggerated, they must be converted at a rate that results from reducing lender capital (capital: paid-in cash investments plus retained earnings). Therefore, rating agencies must examine the lender portfolio prepayment rate assumption, as those lenders will experience a decrease in capital if those assumptions prove false. Any significant loss of lender's capital always implies a reduction in its credit rating. Credit rating agencies will be the primary beneficiaries and users of the present invention.
[0022]
For investment financiers, the system of the present invention establishes a standardized prepayment methodology that allows merger and acquisition advisors to be able to quantitatively measure balance sheet risk at target banking or mortgage companies. I do. Furthermore, the use of the investment financiers of the present invention includes applications to bond security. Securitization means that a pool of mortgages or other debt securities is purchased by investment financiers (within their capacity as underwriters) and resold to institutions and public investors as restructured securities The method will be described. Typically, these security benefits the debt originator. Because they achieve a significant acceleration in realized gains, and because they significantly diversify risk by selling important aspects of bonds to asset underwriters and the like. However, typical bond securitization proceeds with a parent lender that holds significant prepayment risk. If the prepayment rate accelerates beyond the level assumed in the security pricing process, the parent lender is liable. Accordingly, the present invention improves upon and makes the security method more efficient by measuring the prepayment behavior and scoring expected by accepted industry standard methods. Again, this reduces costs for all participants and allows more capital to be used for low cost consumer borrowing.
[0023]
For investors, the method of the present invention may be a method of making an investment decision based on the risk of prepayment of debt securities quantified by lenders to which the investor may wish to invest, or a method of making individual bond To provide a method for assessing the relative stability of mortgage securities supported by.
[0024]
These and other advantages of the present invention are described in the detailed description below.
[0025]
(Detailed description of preferred embodiments)
FIG. 1 shows an outline of the method according to the present invention. The mortgage broker or lender first receives a loan application from borrower 10. That information is transmitted electronically to the present invention which analyzes the loan application information 12 and categorizes it into various categories related to scoring potential loans. The loan application is analyzed based on the information needs of the sophisticated mathematical model resident of the present invention. A prepayment score is then derived 14 for the particular consumer in view of the interest rate environment in which the loan is being processed (ie, rising or falling interest rates) as a function of the particular loan type being requested. As mentioned earlier, this score is indicative of a particular consumer's propensity to prepay. Then, the advanced repayment score is returned to the lender 16. The lender can then create a customized loan product that rewards the consumer's favorable prepayment behavior 18.
[0026]
FIG. 2 shows an outline of a system according to the present invention. Loan originator 20 receives offers from potential consumers. The offer is then entered into the loan originator's data transport channel 22. Such data transport channels 22 include (but are not limited to) email, fax, the Internet, and common electronic means. Other loan originators 34 also send their respective consumer offers on their own data transport channels 36.
[0027]
In the present invention, the loan application 24 may be carried by other digital electronic means such as the Internet 28 or a wireless communication method. An electronic loan application 40 is entered into the system of the present invention via a communication server 42. The loan information for a given consumer is then submitted to the application parser 52. The application parser 52 divides the information into loan information 58 and applicant information 56. Loan information 58 is information about the amount, terms, down payment, type of loan, and other significant and amount to be loaned. Applicant information 56 is information such as name, address, social security number, and other demographic information about the applicant.
[0028]
The loan information 56 is supplied to a prepayment model library database 66. This advanced repayment model library database 66 includes information on advanced repayment history data 62. The results are provided to a model training server 64, which processes the prepayment history data 62 for both each individual and demographic group, and updates the model library database 66. Once the loan information 58 has been processed by the prepayment model library database 66, an analytical prepayment model 60 (based on the loan information 58) is provided to the prepayment calculation server 46. The prepayment calculation server 46 obtains additional information from an econometric model 48 that establishes a wide variety of variable relationships. The econometric model 48 generates chronologically arranged interest rates, mortgage rates and other economic parameters, including the scenarios utilized by the prepayment calculation server. These scenarios are generated from raw discrepancy sequence (LDS) logic rather than using random number generation. LDS logic provides significantly higher model accuracy with the same number of scenarios.
[0029]
Once the prepayment score 44 is obtained by the prepayment calculation server 46, the prepayment score 44 is then sent to the communication server 42 and the Internet (or other electronic channel) 28 via the data transport channel 22 or 36. Is sent to the loan originator 20 or 34, which can approve or deny or create a customized loan product for the consumer.
[0030]
The advanced repayment score 38 is calculated based on the following model. The unique prepayment analysis of the present invention is conceptually shown as follows. The following variables
(Equation 9)

Is a vector of applicant data and loan parameters.
(Equation 10)

Are a series of econometric scenarios based on the Raw Discrepancy Sequence (LDS), which are generated by an RTH-linked exponential econometric model. Thus, the model is a set of stochastic differential equations that describe the dynamics and interactions of the key macroeconomic indicators, each of which is associated with a prepayment propensity calculation.
[0031]
Analytical prepayment model
[Outside 3]

(Depending on the type of loan) is a repayment evaluation value p in a predetermined scenario based on the applicant's data (A), loan parameters (L) and econometric parameters (E). _s Is adjusted to calculate
(Equation 11)

[0032]
The total prepayment accumulated up to time T in scenario s can be calculated as follows:
(Equation 12)

[0033]
Then, the total prepayment at time T is given by:
(Equation 13)

[0034]
And finally, the prepayment score is as follows.
[Equation 14]

[0035]
The analytic model that generates the prepayment score is further formed based on later studies, relating to additional external behavior or by econometric factors, and also by scoring the aforementioned behavior of the mortgage broker behavior. Is also good.
[0036]
The invention may be represented in alternative embodiments in the form of a credit engineering workstation (hereinafter CEW). This CEW (described more fully below) provides a user interface that allows the loan originator to perform all prepayment calculations, model analysis, and the presently described pricing of the present invention using said first prepayment model. Have.
[0037]
The CEW operates in a Unix (R) or Windows (R) NT (Windows (R)) environment using Oracle, SQL Server, Sybase, DB2 or Infomix database support. The CEW uses CORBA, or a structured object model, as well as a JAVA / HTML browser-based graphical user interface.
[0038]
All of the CEW subroutines contribute to the ultimate goal of determining a consumer's prepayment propensity. For example, the subroutine of the present invention provides a method for generating various interest rate scenarios and a subsequent economic scenario model fitting method that fits the modeled interest rate scenarios to historical and current interest rate yield curve performance and other macroeconomic indicators. to support.
[0039]
Portions of this system include reward pricing logic for efficiently measuring and pricing reward impact on consumer prepayment behavior. For example, it is most beneficial to lenders who reward consumers for not prepaying the lender's loan. Such rewards may be assessed for the impact of consumer prepayment. Thus, the system allows end users to design a formal reward structure and test the impact on expected consumer prepayment behavior.
[0040]
Various user-definable screens also establish default spreads, prepayment spreads, broker commission schedules, and other financial factors that affect the pricing of goods to be offered to consumers. Various other economic scenarios are gathered by the user interface and combined with various prospective and default data as well as other lender-defined baseline results in reasonably priced end-user mortgage agreements.
[0041]
FIG. 3 shows further information regarding the CEW of the present invention. The system includes a user interface module 70, which is a basic graphical user interface, and other software that allows the originator to provide information about the consumer who wants to borrow money from the lender. This user interface module allows a collection of loan attributes 76, applicant attributes 74 and reward program attributes 72. In addition, user interface module 70 collects or calculates spreads, broker commissions and other costs associated with loan 78. Loan attributes 76 and other loan-related costs, along with other information, are provided to a pricing engine 84 that helps generate an appropriate loan price 86.
[0042]
A loan attribute 76, an applicant attribute 74, and a reward program attribute 72, all of which affect the value of the loan, are supplied to the prepayment calculation server 80. The prepayment calculation server 80 receives input from various prepayment model parameters and generates a prepayment score 82.
[0043]
FIG. 4 is a block diagram showing the interaction required to generate a prepayment model. Consumer information 96 comprising applicant attributes 74 and loan attributes 76 is provided to a prepayment model fitting 92 module. The advanced repayment model fitting 92 establishes various advanced repayment model parameters 94 based on the advanced repayment history data 90. Once the appropriate prepayment model has been generated by the prepayment model fitting 92, the model may be used to calculate the prepayment score for a given particular consumer as required by the type of loan to the consumer. Returned to the compute server. The prepayment server also benefits from input from the econometric model scenario generator.
[0044]
FIG. 5 shows the interaction for the econometric model. Econometric model scenario generator 106 receives inputs from econometric model fitting module 104 and LDS scenario 108. The econometric model fitting module 104 obtains information from the econometric history data 100 and the current market environment 102, including, but not limited to, information on rising and falling interest rates and their trends. . The information from econometric history data 100 relates to the demographic group to which the consumer belongs and other econometric information such as age, income, credit rating, occupation and other factors. Information from the current market environment 102 relates to the direction and rate of change in interest rates. The econometric model scenario generator 106 processes the information and generates various scenarios based on the information.
[0045]
Returning to FIG. 3 again, the prepayment calculation server 80 generates the prepayment score 44 for the particular consumer. The prepayment score 44 is based on the established prepayment model and its generated econometric model. The advanced repayment score 44 is transmitted to the pricing engine 82 to determine the price of the loan product presented to the consumer.
[0046]
FIG. 6 shows additional parameters used by the user interface module to generate various scenarios. An additional aspect of the invention provides for the creation of a new product. Strategy optimizer 122 is based on the consumer's underwriting of the offered product, and inputs to and from other products are placed on the market. Strategy optimizer 122 generates a marketing plan based on the individual lender's portfolio. Such marketing plans have been able to help provide new products to markets that are more beneficial to lenders. The system includes an optimizer 124 that provides a way to provide loan products to consumers with the most favorable prepayment characteristics, ie, low propensity to prepay. The system also includes a loyalty optimizer 126 that models and defines offers to consumers and other incentives to reward financially favorable consumer behavior. Channel optimizer 128 is part of the present invention. The channel optimizer 128 analyzes the transport channels of the financial product offering to evaluate and determine the channels that are the most efficient way to transport various financial products. This system also includes a database optimizer 130 that receives and organizes information from various databases to continuously produce sophisticated prepayment history data 90 and econometric history data 100.
[0047]
The target platform on which the system of the present invention runs may be any system based on an Intel Pentium processor with 32 megabytes of RAM, hard disk storage and retrieval, and communication capabilities using the TCP / IP protocol. Is. Alternatively, the system also runs under the UNIX operating system on the Sun Solaris platform. In any case, the display for the user is expected to be capable of outputting a hardcopy report. In a typical operation, users away from the system site access the system via a private network or the Internet and send the information necessary for the present invention to make the desired calculations leading to a prepayment score. . This score is then sent to the requesting user at the remote terminal.
[0048]
A system and method for prepayment score generation has been described. Those skilled in the art will recognize that other modifications of the present invention are possible without departing from the scope of the invention.
[Brief description of the drawings]
FIG. 1 shows an overview of the method according to the invention.
FIG. 2 is a block diagram of the present invention.
FIG. 3 is a block diagram showing a user interface module connection.
FIG. 4 is a block diagram showing an interaction with advanced repayment history data.
FIG. 5 is a block diagram illustrating an interaction with an econometric model.
FIG. 6 is a block diagram illustrating factors used by a user interface module.

Claims (16)

A system for determining a consumer's prepayment propensity, comprising: a plurality of loan origination terminals for receiving and transmitting a consumer mortgage loan application;
A network connected to the plurality of loan origination terminals for receiving the transmitted consumer mortgage loan application;
A communication server connected to the network for receiving the transmitted consumer mortgage loan application;
An application parser connected to the communication server for receiving the consumer mortgage loan application transmitted from the communication server and analyzing the information into loan information and applicant information;
A prepayment model library database including a loan payment model connected to the application parser for receiving the loan information and fitting the loan information to a loan payment model and transmitting a loan model consistent with the loan information. ,
The advance having the advanced repayment score generation model connected to the advanced repayment model library database for receiving the loan model and calculating an advanced repayment score based on the loan payment model and the advanced repayment score generation model. A repayment calculation server, wherein the prepayment calculation server is further used for transmitting the prepayment score to any one of the plurality of loan original terminals via the communication server and the network. . The system for determining a consumer's prepayment propensity according to claim 1, wherein the prepayment model library database further comprises:
A model training server for generating a loan payment model for the prepayment model library database;
Prepayment history data connected to the model training server, wherein the prepayment history data further includes prepayment statistics for various types of loans. 2. A system for determining a consumer's prepayment propensity according to claim 1, wherein said prepayment calculation server further comprises: loading an econometric parameter to the prepayment calculation server. It has an econometric model that generates scenarios based on the Crepancy sequence (LDS). 2. The system for determining a consumer's prepayment propensity according to claim 1, wherein the prepayment score is calculated by the following formula.

Here, T is time and P is early repayment. 5. The system for determining a consumer's prepayment propensity according to claim 4, wherein the sum of prepayments at time T is calculated by the following equation.

Here, S is the number of scenarios, and P is the prepayment for a given scenario. A system for determining a consumer's prepayment propensity according to claim 5, wherein the sum of prepayments accumulated up to the time in scenario s is calculated by the following formula.

Here, p (t) is an advanced repayment value. 7. The system for determining a consumer's prepayment propensity according to claim 6, wherein the prepayment value for a given scenario is calculated by the following formula:

Where A is the applicant data, L is the loan parameter,
[Outside 1]

Is an analytical prepayment model. A method of determining a consumer's prepayment propensity, comprising:
Collecting loan and applicant information with the loan originator,
Transmitting the loan and applicant information over a network;
Including receiving the loan and applicant information at a service bureau,
The service bureau computes an expected prepayment propensity score for the applicant, and the service bureau returns the prepayment propensity score to the loan originator over the network. 9. The method of determining a consumer's prepayment propensity according to claim 8, wherein calculating an expected prepayment propensity score for the applicant comprises parsing the information into loan information and applicant information. including. The method of determining a consumer's prepayment propensity according to claim 9, further comprising providing the applicant information to a prepayment model library database and providing the loan information to a prepayment calculation server. Including. 11. The method of determining a consumer's prepayment propensity according to claim 10, further comprising: determining the prepayment model that best fits the loan information, and providing the model to the prepayment calculation server. Has a prepayment model library. 12. The method of determining a consumer's prepayment propensity according to claim 11, further comprising the prepayment calculation server receiving a prepayment model and an econometric model, wherein the prepayment calculation server further comprises: Calculate the prepayment propensity score of the applicant. 13. The method of determining a consumer's prepayment propensity according to claim 12, wherein the prepayment propensity score is calculated based on the following formula.

Here, T represents time, and P represents advanced repayment. 14. The method for determining a consumer's prepayment propensity according to claim 13, wherein the total prepayment at time T is represented by the following equation.

Here, S represents the number of scenarios, and P represents the advanced repayment amount of a given scenario. The method of determining a consumer's prepayment propensity according to claim 14, wherein the sum of prepayments accumulated up to the time in scenario s is calculated by the following formula.

Here, p (t) is an advanced repayment value. 16. The method of determining a consumer's prepayment propensity according to claim 15, wherein the prepayment value for a given scenario is calculated by the following formula:

Where A is the applicant data, L is the loan parameter,
[Outside 2]

Is an analytical prepayment model.

Images