The Activity in the Danish Money Market

Literature

Damm, Birgitte and Anne Reinhold Pedersen (1997), New Money-Market Statistics, Danmarks Nationalbank, Monetary Review, 3rd Quarter.

Danmarks Nationalbank (2003), Monetary Policy in Denmark.

Danmarks Nationalbank (2004), Financial stability.

Furfine, Craig H. (1999), The Microstructure of the Federal Funds Market, Financial Markets, Institutions and Instruments, Vol. 8(5), December.

Furfine, Craig H. (2003), Interbank Exposures: Quantifying the Risk of Contagion, Journal of Money, Credit and Banking, Vol. 35(1), February.

Millard, Stephen and Marco Polenghi (2004), The relationship between the overnight interbank unsecured loan market and the CHAPS Sterling system, Bank of England Quarterly Bulletin, Spring.

Pedersen, Anders Mølgaard and Michael Sand (2002), The Danish Money Market, Danmarks Nationalbank, Monetary Review, 2nd Quarter.

Appendix: Method of identification of uncollateralised day-to-day money-market loans

Payments in kroner among the monetary-policy counterparties take place via Danmarks Nationalbank's payment system, Kronos, and are settled via accounts with Danmarks Nationalbank. Thus, trading in uncollateralised day-to-day loans among the monetary-policy counterparties may be identified by using an algorithm searching through the individual payments in Kronos. An equivalent analysis method has been used previously by others. Furfine (1999) analysed the structures in the US day-to-day money market (the federal funds market) based on data from a search algorithm with assumptions similar to those in Danmarks Nationalbank's algorithm, while Millard and Polenghi (2004) carried out a similar study of the money market in the UK. Furfine (2003) used the same method and data to estimate systemic risks in the US interbank market. Danmarks Nationalbank (2004) contains a similar analysis.

The idea of the search algorithm is simple. For instance, if bank A grants bank B an overnight loan of kr. 100 million at an annual interest rate of 2 per cent, a payment of kr. 100 million will be made from bank A's current account with Danmarks Nationalbank on the day of the agreement, while on the same day a payment of kr. 100 million will be made to bank B's current account. The loan of kr. 100 million is repaid from bank B to bank A on the following day with accrued interest of kr. 5,555.56 (= 0.02*100,000,000/360). The algorithm runs through all payments via current accounts of banks and mortgage-credit institutes and searches for repayments on the following banking day "matching" the accrued interest.

The agreed interest rate for the individual loans is not known in advance and may depend on the counterparty's credit rating, the time of transaction and liquidity in the money market in general. Thus, the algorithm searches for counterbalancing payments where the accrued interest may vary within a certain interval. This interval is determined on the basis of day-to-day interest rates that are reported daily to Danmarks Nationalbank from 13 banks. The lower limit of the interval is determined by the lowest reported interest rate for uncollateralised overnight or tomorrow-next transactions less 0.25 per cent per annum. Similarly, the upper limit of the interval is determined by the highest reported interest rate plus 0.25 per cent per annum. The interval is not constant as the difference between the lowest and the highest reported interest rate varies from day to day. The interest-rate interval is shown in Chart 3. 

The reason for adding and deducting 0.25 per cent is that the reported interest rates are based on a weighted average of all trades in the relevant segment. The interest rates for the individual trades may therefore deviate from the reported average. Since the algorithm does not differentiate between overnight and tomorrow-next transactions^[6], the lowest and highest, respectively, of the two reported interest rates is used as the point of departure. To avoid that other types of day-to-day transactions, cf. Box 1, than uncollateralised day-to-day loans are identified by the algorithm only payments of at least kr. 1 million and which are in whole hundred thousands (i.e. ending in five zeros) are included. A payment of kr. 2.1 million is thus included as a potential loan, while a payment of kr. 2.15 million is not. The reason for this criterion is that "round figures" are the market standard for deposit transactions. Furthermore, it ensures that foreign-exchange swaps are not included, as these are typically agreed in round amounts in dollars which predominantly result in odd amounts in kroner. Collateralised trades in the form of repos are settled by VP Securities Services and thus not via individual transactions between current accounts. Loans between institutions that are both part of the same group and that both have current accounts with Danmarks Nationalbank are disregarded.

In the 4th quarter of 2003 and the 1st quarter of 2004 approximately 323,000 payments were made in Kronos between current accounts equivalent to 2,561 payments per banking day, cf. Table 5. The algorithm identified approximately 21 per cent of these payments as candidates for uncollateralised day-to-day money-market loans, and matching repayments of around 7 per cent hereof were found.

Identification of uncollateralised day-to-day loans		Table 5
	Number	Value (kr. million)
Payments per day in Kronos	2,561	209,331
Of which potential day-to-day loans	544	100,255
Of which identified day-to-day loans	39	9,305
Note:    Payments are here defined as a payment from one current account to another. Source:  Danmarks Nationalbank.

The loans identified by the algorithm are solely uncollateralised day-to-day loans between monetary-policy counterparties, i.e. banks or mortgage-credit institutes with interest-bearing current accounts with Danmarks Nationalbank. Consequently, only parts of the market for uncollateralised day-to-day money-market loans in kroner are included. The analysis does not comprise trading with foreign institutions and with small institutions that are not monetary-policy counterparties.