First of all, we would like to point out that the results of the EBA's empirical analysis across asset classes confirm that some covered bonds achieve the same average ranking as government bonds. We agree with this analysis and see no reason to discriminate against covered bonds based on other criteria. If this finding is taken into account when the Commission decide on the final definition of the LCR in June 2014, Denmark will no longer be classified as a country with insufficient liquid assets. If not we agree that the Danish krone should be on the list of currencies with constraints on the availability of liquid assets.
Concretely, we strongly agree that, for markets to remain liquid, it is necessary to have a significant portion of assets not locked up and that this need is relatively greater for small markets than for large ones. The Danish gov-ernment bond market is both by international comparison – as well as compared to the Danish covered bonds market – small due to the low level of government debt.
Requiring financial institutions to hold a large proportion of the government bonds in circulation in a liquidity buffer would in itself negatively impact the liquidity of these bonds, especially considering the already substantial level of estimated buy-and-hold investors in the market. Monetary and fi-nancial institutions have historically been minor holders of Danish government debt.
In addition, the majority of Danish government bonds outstanding are long-term bonds, not well suited as liquidity risk management instruments. This reflects that it for years has been government debt management policy to have low refinancing risk (the average time to maturity and duration of the government bond market are among the highest of the OECD countries). The amount of long term bonds may thus be considered an additional type of locked-up" assets.
Generally, a discrimination against assets of otherwise comparable liquidity properties thus creates incentives to issue more government debt as well as to shorten its maturity structure.
We therefore see several challenges in setting a buffer of free-floating assets – unless the buffer is either based on calculations on the supply side or is very large. There is thus no reason to assume that setting a buffer of free-floating assets as a fixed portion of demand for liquid assets ensures liquidi-ty on the government bond market. Setting a very high buffer may however reduce the risk of the government bond market becoming illiquid.
Firstly, to maintain liquidity as it is today the buffer would have to be larger than 50 per cent as opposed to the 25 per cent suggested by EBA.
Secondly, the factors involved in calculating the need for a buffer are vola-tile. As one example, the market value of government bonds is about DKK 90 bn. lower today compared to September 2012 – mainly due to higher in-terest rates. The lower market value of government debt today as well as a (slightly) higher share of buy and hold investors has – all else equal – in-creased the shortage of liquid assets from 2 per cent in September 2012 to about 16 per cent.
Going forward, changes in interest rate levels may significantly change the availability of liquid assets due to changes in the market value of govern-ment debt. Volatility should also be reflected in the size of the buffer to en-sure liquid government bond markets."
Overall, we agree with the approach of excluding locked-up assets from the pool of available liquid assets as well as the assumptions on the specific amounts of locked up assets in the Danish insurance and pensions sector and with foreign investors, cf. Q1.
As noted in Q1, the majority of Danish government bonds outstanding are long term bonds, not well suited as liquidity risk management instruments. This further limits the amount of available liquid assets that can be used for banks' liquidity management. The amount of long term bonds may thus be considered an additional type of locked-up" assets."