Settling Low Value Payments in Real Time: RT G S

Transcription

1 Settling Lw Value Payments in Real Time: The experience f the Mexican RT G S Biliana Alexandrva Kabadjva Francisc Slís Rbleda Speaker: Biliana Alexandrva Kabadjva August 2010 * The views and cnclusins presented in this presentatin are exclusively the respnsibility f the authrs and d nt necessarily reflect thse f Banc de Méxic. Intrductin Outline Perfrmance measures Results Final Remarks 1

2 Intrductin Payment systems have evlved ver time; Changes driven by technlgy are bserved in many aspects f the payment service industry; Onging innvatin is likely t diversify payment types cmpeting n cnsumer service level, whereas efficiency and cst reductin culd be the main reasns fr integratin f payments prcessing and settlement; It culd be that in a near future real time high-value payments, lw value electrnic payments and even card payments will be settled tgether; th T achieve this, settlement engines need t incrprate a Liquidity Saving Mechanism, which allws settlement f a large number f lw value payments with a minimum pressure n liquidity cnsumptin. Intrductin Swiss Interbank Clearing System, SIC, is the best knwn example f a Large Value Payment System (LVPS) that prcesses lw value payments; Nevertheless, ther cuntries like Czech Republic, Serbia, Slvakia, Turkey, Ukraine, Saudi Arabia and Mexic, which als use ne system t settle whlesale payments and retail payments tgether; In the case f Mexic is the Real Time Settlement (RTS) Payment System, SPEI, which settle tgether high-value and lw- value payments It started peratin n the 13th f August 2004 in rder t settle Large Value Payments; SPEI has integrated a Liquidity Saving Mechanism (LSM), which allws cntinuus netting f payments during the day; 2

3 The Mexican Case SPEI settles payment rders n real time and its peratr, the Mexican Central Bank, charges less than a 0.05 USD per payment; SPEI settles n average arund 300,000 peratins daily. Mre than 80% f the transactins are payments with value lwer than 10,000 USD and, nly 1.3% f the transactins are abve 1,000,000 USD Anther way t send electrnic transactins in Mexic is thrugh Fund Electrnic Transfer (TEF), which uses differed settlement (t+1) and is perated by Cecban, S.A. de C.V. a cmpany wned by banks; Electrnic Transactins Number f peratins TEF SPEI

4 The study The present study evaluates the effect f settling in real time a large number f lw-value payment transactins n the liquidity perfrmance f SPEI when prcessing high-value payments; T that end, we create an artificial envirnment, in which we use histrical data fr 30 days and reprduce the peratinal cnditins f SPEI; Fr each f these days, we structure the transactins in fur sets, delimited accrding t their value: Set ne: all payments; Set tw: subset f payments with value higher than 100, MXN*; Set three: subset f payments with value higher than 1,000, MXN; Set fur: subset f payments with value higher than 10,000, MXN; We measure tw aspects f the perfrmance f SPEI in tw scenaris (described later): required participants funds (liquidity cnsumptin) and speed f settlement (measure in terms f amunt paid multiplied by the time delay). *The values f the lwer bund f each subset f payments are chsen as a reference, withut fllwing any specific cnventin. The current exchange rates f the Mexican pes t an US dllar and t an Eur are 13.10MXN = 1USD and 16.70MXN = 1Eur respectively. Operatinal rules f SPEI SPEI receives payment instructins cntinuusly during the day, which are placed in a queue. Operatin starts at 8:30 and clses at 17:00. During peratin time, a settlement prcess (SP) is executed at the latest 20 secnds after receiving a new payment. Payment instructins nt settled in a certain SP are kept in the queue and are cnsidered fr settlement in the subsequent prcesses. After executin f the last SP f the day, payments in the queue are canceled. 4

5 Outline Intrductin Perfrmance measures Results Final Remarks Measuring the speed f settlement Ntatin P is the set f payments received by SPEI in ne day; I is the set f participants in SPEI; pij is a payment in P, frm participant i t participant j; t p is the number f minutes frm the first SP launched immediately after the receptin f the payment instructin p, until its settlement; ξp is the amunt in MXP f p. 5

6 Measuring the speed f settlement Then the delay V, which measures the speed f payments settlement during ne day, is defined n the fllwing way: V = t p ξp Nt Nte: if liquidityidit is sufficient i t settle every pij P in the next SP launched immediately after the receptin f pij, then V =0. Measuring liquidity cnsumptin In Large Value Payment Systems, the term liquidity is used t define the funds that the participants have t cver their payment bligatins; Thse funds cme primary frm tw surces: Participants resurces frm previus balances r received as electrnic transactins frm ther payment systems and, Payments received during the day frm the rest f the participants i t Fr ur study, the measure f liquidity cnsumptin is related nly t participants resurces 6

7 Measuring liquidity cnsumptin A is the set f accunts f the direct participants in SPEI. la is the levell f wn resurces used in the accunt a during ne day; i I the payment rders sent are presented as: P Penv= j II pij whereas the received payments are dented as: Prec= j I pji Measuring liquidity cnsumptin l min a is the lwer required level f liquidity fr settlement during the teday, which is determined ed as fllws: s l min a = min{(penv Prec), 0} l a max is the required level t have V =0. Therefre l min a la is a necessary cnditin in rder t settle all payments py with la. In additin fr any la l a max, V =0. 7

8 Measuring liquidity cnsumptin Then, L is the measure n the macr level f liquidity cnsumptin, defined as: L= a A la p P ξp This measure is applied here t the histrical data tested in a simulated envirnment. Outline Intrductin Defining perfrmance measures Results Final Remarks 8

9 The study We created an artificial envirnment, in which we use histrical data frm 30 days and reprduce the peratinal cnditins f SPEI. We nly included d payments received in SPEI frm 8:30 t 17:00 Fr each f thse days we structure the transactins in fur sets, delimited accrding t their value: Set ne: all payments; Set tw: subset f payments with value higher than 100, MXN*; Set three: subset f payments with value higher than 1,000, MXN; Set fur: subset f payments with value higher than 10,000, MXN; *The values f the lwer bund f each subset f payments are chsen as a reference, withut fllwing any specific cnventin. The current exchange rates f the Mexican pes t an US dllar and t an Eur are 13.10MXN = 1USD and 16.70MXN = 1Eur respectively. The study In the first scenari: fr each set f payments we btain Lmin, the minimum amunt fr participants funds fr which all payments are settled; Under this cnditin we calculate the delay V. In the secnd scenari: fr each set f payments, we calculate Lmax; In this case delay V = 0. Fr each scenari we cmpare the perfrmance f SPEI with and withut the settlement f lw-value payments. 9

10 Histrical Data Number f SP Payments with Value > Payments with Value Payments with Value Day executins All payments 100,000 > 1,000,000 > 10,000, ,637 31,280 7,116 2, ,578 33,211 7,563 2, ,326 41,367 8,559 2, ,695 33,616 8,556 2, ,899 31,728 8,140 2, ,436 36,093 8,219 2, ,731 38,203 8,481 2, ,347 43,045 8,876 2, ,322 36,703 9,363 2, ,098 30,611 7,804 2, ,978 30,525 7,464 2, ,068 35,190 9,041 3, ,774 42,249 8,861 2, ,330 34,711 8,760 3, ,009 32,532 8,020 2, ,682 35,879 8,562 2, ,387 44,756 10,374 3, ,860 59,374 12,853 4, ,294 31,954 8,601 2, ,136 30,388 7,952 2, ,316 39,758 9,845 2, ,648 34,747 8,706 3, ,832 43,507 10,082 3,665 Liquidity cnsumptin Payments > Payments > Payments > All Payments 100,000 1,000,000 10,000,000 Average Lmin Variance Est. Div Average Lmax Variance Est. Div

11 Settlement delay* Payments >100, Payments >1,000,000 Payments >10,000,000 Settled with all payments Settled separately Settled with all payments Settled separately Settled with all payments Settled separately V Average 11,486, ,636, ,178, ,198, ,138, ,099, Est. Div. 3,330, ,327, ,496, ,591, ,493, ,321, *The settlement delay is presented in terms f 1,000, Mexican pess in ne minute Lw Level Liquidity Cnsumptin Delay f Payments > 100, Delay v with all payments nly payments > 100,

12 Lw Level Liquidity Cnsumptin Delay f Payments > 1,000, Delay v with all payments nly payments > 1,000, Lw Level Liquidity Cnsumptin Delay f Payments > 10,000, Delay v with all payments nly payments > 10,000,

13 Lw level f liquidity cnsumptin Lmin Lmin All Payments Payments > 100,000 Payments > 1,000,000 Payments > 10,000, High level f liquidity cnsumptin Lmax, V= Lmax All payments Payments > 100,000 Payments > 1,000,000 Payments > 10,000,

14 Outline Intrductin Perfrmance measures Results Final Remarks Final Remarks This study has allwed us t evaluate the speed f settlement, measured thrugh delay V and the liquidity cnsumptin, measured thrugh L in tw cases: Lw liquidity cnsumptin, High liquidity cnsumptin withut delay. Accrding t ur results, lw value payments d nt have a negative impact ver the speed and liquidity cnsumptin; One pssible extensin t the present wrk is t find the cnditins and the peratinal rules under which a smaller delay (V) is btained given a specific balance 14