I have been puzzled for some time by the loose use of the term "reserves" when talking about banking. I hear people talking about the deposit multiplier (which itself is a myth, see my post To Lend or Not To Lend) as if the deposits taken from retail customers form part of the capital against which banks - supposedly - lend. Er, no, they don't. They form part of the CASH reserves that banks hold to support settlement of lending. Bank capital has a totally different composition.
Let me explain.
Cash reserves are the MONEY the bank has available in its cash reserve account at the Bank of England at the end of each day. They are intended to cover the expected physical cash drawings of loans and deposits over the next day. In practice most retail banks don't have sufficient cash from deposits to cover drawings for the next day, so they borrow the money either from other banks (usually investment banks) through the interbank lending market or directly from the Bank of England.
Capital is the EQUITY of the bank. Under the accords agreed by the major financial centres of the world in Basel, it is divided into different tiers:
- Tier 1, which usually consists of retained profits and proceeds from share issues (common stock). This capital is FULLY at risk in the event of the bank becoming insolvent
- Tier 2, which consists of subordinated debt (debt which is repaid only after all other claims have been met, so might not be repaid at all), preference shares and other instruments that can be treated as equity (i.e. not repaid in the event of losses), general provisions (money the bank sets aside to cover loans it believes will not be repaid), revaluation reserves (money acquired as a result in change of value of an asset, for example through increasing value of property) and undisclosed reserves (retained earnings that for some reason have not been disclosed in the published accounts). This capital is also at risk in the event of the bank becoming insolvent, but only to the extent that Tier 1 capital does not cover the losses.
- There is also a Tier 3, which consists of subordinated debt used only as a buffer against market risk losses. This does not form part of what are loosely described as "capital reserves" supporting lending.
Capital is the bank's "cushion" against possible losses, and it is only meaningful to talk about it in relation to bank debt. This is unfortunate, because most of the recent writing on "capital reserve requirements" talks about it in relation to bank assets, which causes enormous confusion. Bank debt - money the bank has borrowed - mainly consists of retail deposits and issued securities. The larger the proportion of capital in relation to debt, the more resources the bank has available to call upon to settle losses and protect depositors and bondholders without having to call for assistance from taxpayers or seek a merger with a richer competitor (as Midland did when it invited HSBC to take it over). Ideally banks that lend under the fractional reserve system - where they don't have to have 100% cash reserves available to cover their lending - need to have a reasonably substantial equity cushion available to protect depositors. In practice banks have over the last few years reduced their equity cushions to tiny proportions and relied more and more on FSCS insurance and, as a last resort, taxpayers to guarantee depositors and bondholders in the event of losses. It is this imbalance that the Basel committee, and the UK's Independent Commission on Banking (ICB) attempted to address by raising capital requirements.
When the ICB produced its report recommending changes to banking structure, it suggested two changes to capital requirements. Note that these are quoted in relation to assets not liabilities, despite my comments above:
- that systemically-important banks (the ones that are "too big to fail") should have sufficent capital to cover at least 10% of risk weighted assets
- additionally, that banks that do both retail and investment banking should ringfence their retail operations in a separate legal entity with its own capital covering at least 10% of risk weighted assets belonging to those retail operations.
These requirements have NOTHING to do with cash reserves or bank deposits. But confusion is widespread. Even the estimable FT Alphaville blog got this wrong when it pointed out that retail banks don't have sufficient deposits to cover 10% of their existing lending. They aren't being asked to hold sufficient deposits to cover this. Deposits are debt - not subordinated debt, just general debt - so don't form part of a bank's capital. Really, the banks are being asked to increase their retained profits and/or issue more shares (or debt instruments that can become shares if necessary).
Now, I pointed out in a previous blogpost that the increasing spread between the awful rates paid to savers and the exorbitant rates charged to borrowers could be explained by the banks' need to increase their capital. Various people argued with me about this, but I stand by my remarks. Banks make money by lending at higher rates of interest than the rates they pay to depositors, and that difference (once other costs have been met) is either paid out in dividends to shareholders or is added to retained profits in Tier 1 capital. The wider that difference, the more banks earn and the more they can retain while still maintaining shareholder dividends. Increasing capital requirements is the best excuse you could possibly give the banks for ripping off both savers and borrowers.
I have read blogposts that argue that increasing capital requirements should have no effect on the cost of bank lending - since it is only replacing bank debt with equity - but may make lending harder to obtain. I beg to differ. I think that increasing the capital requirement is unlikely to affect availability of lending, and certainly is unlikely to improve the risk profile and management of lending. But it will increase the cost. It generally costs banks more to raise equity capital than it does to borrow on the interbank markets or issue bonds, so forcing banks in effect to replace debt with equity will raise their cost of capital and they are likely to pass this on to their customers. And diluting shareholders' capital by issuing new shares almost always depresses the share price, with consequent impact to existing shareholders. Disguising new share issues as convertibles such as preference shares doesn't fool the markets. Increasing capital by rights issues or new convertible debt issues is going to be unpopular with shareholders and institutional investors, who in the end call the shots - after all, they control the remuneration of bank directors. Small wonder then that banks prefer to increase retained profits by raising borrowing rates and depressing savings rates. It's a no-brainer, really.