Improving Overnight Loan Identification in Payments Systems

Abstract

Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in RésuméIl y a peu d'information au sujet de la répartition des actifs et de la formation des prix sur les marchés de gré à gré. En 1999, Furfine a conçu un nouvel algorithme qui fournit des données relativement aux transactions effectuées sur le marché du financement interbancaire. La validité de ces données n'est cependant pas bien établie. En procédant à une permutation des données, j'estime que la limite supérieure du taux de faux positifs quotidien généré par cet algorithme est légèrement au-dessus de 10 %. Je propose des améliorations permettant de réduire cette limite sous 10 % tout en subissant une perte de puissance statistique négligeable. Les résultats donnent à penser que les quantités et les prix des prêts à un jour déduits grâce à cette méthode constituent des estimations valables de l'activité sur le marché du financement interbancaire.

Efficiency and Bargaining Power in the Interbank Loan Market

“…The recent literature on implementation of the Furfine algorithm and associated misclassification error, implies that our estimate of bargaining power can be measured with error. As we did with the efficiency measure, we borrow from Rempel () and model the distribution of type 1 error in the classification of payments into loans.…”

“…The recent literature on implementation of the Furfine algorithm and associated misclassification error, implies that our estimate of bargaining power can be measured with error. As we did with the efficiency measure, we borrow from Rempel (2014) and model the distribution of type 1 error in the classification of payments into loans. 16 Chapman et al (2007) study the bidding behavior of these same participants in daily 4:30 pm auctions for overnight cash and find that, whereas there are persistent violations of best-response functions in these auctions, the average cost of these violations is very small, only a couple of dollars.…”

What drives interbank loans? Evidence from Canada

“…This specialized payments system -e.g., FedWire in the United States and Large Value Transfer System (LVTS) in Canada -which allows participants to quickly process payments of significant size between one another, is the platform used to execute unsecured lending. Rempel (2016) uses a refinement of the Furfine (1999) algorithm to identify overnight loans from the LVTS payments flows. Every unsecured loan consists of a pair of payments exchanged between the same two institutions on successive (working) days, with the interest payment included in the second day's payment.…”

“…Rempel (2016) provides a review of the unsecured lending market in Canada Garriott and Gray (2016). describe the participants and infrastructure underlying the Canadian repo market in greater detail.14 Other short-term funding sources, such as bank certificates of deposit, commercial paper, and securities lending, typically have longer maturities.…”

“…Asset sales in the spot market is another possible means to raise cash; however, settlement conventions in Canada mean that only debt instruments with less than one year to maturity settle on the same day. More pertinently, since asset sales may be triggered by a plethora of motives unrelated to funding needs, we choose to exclude them from our analysis.15Rempel (2016) argues that a first leg with multiple feasible second legs ("non-unique loans") has a higher probability of being falsely identified as an overnight loan, and proposes retaining only such nonunique loans that share sufficient similarities with closest-in-time unique loans.…”