LENDING GOLD FOR LONGER?
THE RISKS AND RETURNS TO MATURITY EXTENSION FOR CENTRAL BANKS
Clifford Smout
Head of Foreign Exchange Division
Bank of England
Slide 1
It is a great pleasure to be here with you today as a member of this panel, particularly since we are here in Istanbul and the panel is chaired by Terry. Last year, in Dubai, I spoke about the UK auctions. This year, Stewart Murray has asked me to discuss aspects of the gold lending market, following on from Hervé Ferhani’s speech yesterday. In doing so, I shall try to keep my comments brief, so that you have plenty of time to question the others on their talks!
Slide 2
I’ll start by saying a few words about the lending market in gold. As Hervé said yesterday, lending by official institutions has increased, from around 900 tonnes in 1990 to 4,830 tonnes last year, according to Gold Fields Mineral Services: these estimates seem entirely plausible to me.
So why has there been this increase? In part, of course, because lending provides a tangible return, of (shall we say) 100bps. On the other hand, it also involves credit risk, and this is important; one of the reasons for holding gold is that it is riskless in credit terms even during a financial meltdown, and this is no longer the case if it is lent out. So much of the return on lending is compensation for credit risk.
Does lending affect the spot price? This is a fascinating topic, which I do not have time to deal with properly today. But I was extremely interested to read a report written recently by Anthony Neuberger, of the London Business School, which was prepared for the World Gold Council. This concluded that the rapid growth of the derivatives market may have led to a somewhat lower spot price, but that the effect was much too small to explain all of the real decline in the gold price seen over the last decade, and was in any case temporary. Professor Neuberger also pointed to the very considerable advantages provided by gold lending to producers, refiners, fabricators and jewellers, in helping to reduce risk, and to holders of gold, by increasing flexibility and returns and thereby making gold a more attractive asset to own.
Slides 3-5
So much for general remarks. What has been happening to the yield curve recently?
Typically, the gold lease curve has looked like this: upwards sloping, but from a fairly low level at short maturities. More recently, though, short rates have backed up, and the curve has flattened. That said, rates remain well below their levels late in September 1999. So what is going on?
Slide 6
The normal upwards slope to the gold yield curve reflects the traditional mismatch between the maturity preferences of borrowers and lenders. Lenders - typically central banks - are concerned about credit risk, and have often refused to lend for more than 6-12 months. This has tended to REDUCE short rates relative to longer rates. By contrast, producers (who account for the bulk of the borrowing) will often wish to borrow long term, to provide price protection for mine development for some years in advance. This has tended to INCREASE longer rates relative to short rates. The bullion banks, who intermediate between the two, take risk in doing so, but the shape of the curve has compensated them for it. Whether the return has been adequate for that risk is another issue!
But recently there have been spikes in rates. Some, for instance in September 1999, fed all the way along the curve, and also affected the spot price. Others, for instance in February and March this year, have been concentrated at the short end. What has changed?
Slide 7
The first, and major, change has been the European central bank agreement on gold (which should NOT be called the Washington or ECB agreement!), signed by 15 central banks in Europe. This provided transparency on central bank intentions on sales, and capped lending at current levels.
What were the effects of the agreement?
Slides 8
Since it was reached, most central banks have reacted in two ways.
First, many (but not all) have lent as much as they are allowed under the Agreement. This has made official lending even less responsive to movements in lease rates than in the past; if rates spike up, these central banks cannot lend any more gold under the Agreement.
Next, many have looked at other ways to increase return. Some have extended the maturity of their lending. I know of one bullion bank whose deposits from central banks of over 12 months in maturity have increased eight-fold in the last two years, admittedly from a low base (though not zero!) Other central banks have looked at products such as lease rate swaps, which extend tenor but with limited credit risk.
Slides 9-11
In making their calculations, central banks tend to look at charts like these. These show risks and returns on lending at different maturities in the past. The risk (measured along the bottom axis) and the return (on the vertical axis) are plotted for various portfolios of deposits from 0 to 6 months. The curve traced out, called an ‘efficient frontier’ by those in the know, shows for each level of risk the best available return.
Now look how returns increase if extensions out to 2 years are allowed. Or out to 5 years. This can be done by using products such as lease rate swaps, with very limited credit risk; similar results would be produced by collateralised lending. It means that for the same risk, returns are noticeably higher. This all looks very interesting to a central bank, or to any other investor in gold.
Slide 12
But I’m a central banker. And that means there is always a but. What are the pros and cons to extending the maturity of gold lending?
Slide 13
First, as we have seen, in the past returns would have been higher if gold had been lent at long maturities. But to be told this after the event is not much of a plus. The question is how far this will continue to be true in the future. For if all central banks follow the same strategy, in a herd, the curve will flatten, as it will if producers choose to borrow less. Already this has happened, and if the curve remains as it has been recently, the additional returns - if any - will be far less than before. Simply put, we cannot conclude that the future will be like the past.
Next, the structural mismatch between short-term lending and long-term borrowing is likely to reduce. This means less risk for the bullion banks, and also less return. I leave it to you and to them to decide whether that is a good or a bad thing.
Third, lending longer means that such gold is unavailable to respond to short-term spikes in rates. Of course, the aggregate amount of gold supplied to the market is the same in both cases, so it is tempting to say this should make no difference. But with regular maturities, a commercial bank which suddenly needs to borrow gold can bid for a new deposit from a central bank, at a rate slightly above that bid by the bank with whom the deposit is maturing. With fewer such maturities, the bank which needs the gold may have to bid more aggressively for deposits, either from the central bank or from its commercial competitors, for instance those who received long-term central bank deposits some time ago. In the long run this should not make much difference. But until we get used to these new dynamics, short rates may well remain volatile.
Slide 14
So what are the implications of this?
First, commercial banks need - more than ever - to follow prudent liquidity policies and be aware of shocks that may cause surges in the demand for borrowed gold. They cannot rely on being able to fund themselves by paying only a little above the going rate.
Second, each individual central bank needs to look at the whole yield curve, when it decides where it wishes to lend. This may involve extending maturities, but it may not. Remember, cash is part of the yield curve. If rates are low but are expected to rise by the central bank, it makes sense for it to withdraw gold from the market, to lend at higher rates later. Taking advantage of such changes in yield curves by trading astutely is exactly what central banks do elsewhere in their investment management operations. Done well, it makes money, and (as in any other market) it also tends to smooth out fluctuations and spikes in rates, at least to some extent. Nevertheless, I expect lease rate volatility to be with us for the foreseeable future.
Slide 15
So what do I conclude?
First, periods of transition are never easy!
Next, the gold lending market is indeed like any other. It responds to demand and supply. Gold is different, but not THAT different.
Finally, much has already changed; short rates have firmed relative to long rates as market participants have altered their views on the relative risks and returns of both.
So predictably enough for a central banker, I conclude that there are both pros and cons to extending gold lending maturities for central banks. Equally predictably, especially for a speaker in the 12.30 slot, I also conclude that in this market at least, there really is no such thing as a free lunch. |
