Banks affect all types of external financing

Effect 1. Investing through FIs can diminish credit risk through (i) deposit insurance and (ii) reduced agency problems and costs associated to the final debtor if his debt-concentration to banks is high.

26 i. If deposit insurance was guaranteed on 100 % of the deposit and if the government issuing the guarantee was default free, depositing money in the bank would be risk-free.

If so, the only interest earned would be an inflation premium and real interest rate. No real interest rate risk premium would be offered, because it is free for the depositor to withdraw the money at any time. Thus, the banks are required to repay the depositor at any point in time. However, the government only cover a part of deposits value if the bank defaults. A government is not default free; there is always a risk that it cannot keep its promise of meeting the deposits (Greek is an example of a country where the credit risk of the government is high). In general, such risk is small. For this reason, the government insurance decreases the risk premium.

Under a financial crisis and given depositors’ perception of deposit risk, their risk preferences might change in favour of less risky investments, because the real credit risk of securities is difficult measure. If prices drop below what the real risk suggests, such drop is irrational. However, it can still be systematic and thus unavoidable. Since security investors obtain knowledge of the real risk via signalling (and because signalling under distress also includes irrational behaviour), the effect of rational risk assessment and irrational risk assessment can be difficult to distinguish. Thus the real price, which the market should correct to over time, is difficult to assess. Even more so as irrational drops can have real economic consequences which indeed must be assessed and factored into the price. A precondition for this argument is that one believes the secondary market to react more severely and more irrationally to the financial crisis (arguments follow later).

ii. The agency problems are expected to increase in the secondary market in parallel to a drop in market value of bonds and equity. Agency problems (i.e. the problem arising when the debtor, here the NFC, chose actions to increase its own utility at the expense of the creditor) increase as the solvency degree of the NFC decreases. Whether the NFC’s incentive to conduct moral hazard is larger for owners than for creditors depends on the owner’s power over the management. In any case, the management will have a larger incentive to make risky decisions and that will increase the risk of default. As market players find signalling less reliable and the probability of agency problems higher, they might look for information sources using less signalling, as for example banks. NFCs usually owe a larger amount to one bank than to any other creditors and therefore the relationship to the bank is more crucial to the NFC. This decreases the NFC’s incentive to create agency problems for banks. Thus, as the perceived risk related to agency problems

27 increases in the security market and the real risk is more difficult to assess, creditors might shift to investments where both problems are reduced, that is in deposits.

As a result of effect (i) and (ii) above, banks’ liquidity improves resulting in easier and cheaper access to finance through banks. Assuming bank’s keep the asset portfolio constant such effect would favour bank-loan finance to NFC’s. The effect is expected to be higher in market-based countries as the systematic risk and hence portfolio volatility is higher.

Effect 2. FI’s activities can themselves create credit risk for the depositors as under the recent crisis when high gearing by the banks combined with risk-sharing resulted in moral hazard by the banks. The higher the default risk of banks, the higher the risk is that depositors do not receive their savings, because deposits are only partly insured. If depositors had underestimated the credit risk of the banks’ assets, depositors withdrew their money when re-assessing the credit-risk. The remaining depositors require a higher liquidity risk premium for keeping money at the saving accounts. The effect would be the exact opposite of Effect 1: Bank-loans will be most affected if banks seek to meet their liabilities by decreasing all assets with the same degree. They do this by selling their securities, stop buying new securities, redeeming loans and stop granting new loans.

Since banks are not the only security traders, but the only bank-loan creditors, the total value of NFCs’ securities in the economy will not decrease as much as the total value of bank loans. One exception is if other security holders also reduce their balance of securities with the same degree as banks. In addition bank loans are affected more, because higher risk premium demanded by depositors is directly transferred to loans, not securities, as other investors than banks set the price in the secondary market.

Effect 3. Liquidity is largely created through lending. If a bank lends money to another bank it receives interest/income. The second bank then lends the money to a third bank and receives interest for this. The third bank does the same thing and so on. Hence the same monetary base shifts hands many times and this is exactly how they all create profit. Some of the money is also lent to NFCs and private debtors, who invest them or spend them. The point is that interbank lending is the main source of short-term capital, and thus it significantly affects liquidity, stability and profit of banks. Lower interbank lending reduces liquidity: The effects are:

(i) Higher transformation risk associated with especially long-term investments such as bank loans. Banks partly use interbank lending as a bouncer if they cannot meet

28 deposit withdrawal and serve the debtors in a specific period. Thus interbank lending is a crucial component in reducing withdrawal risk (and thus liquidity risk) and the complications occurring when matching short-term liquid liabilities with long-term illiquid assets. Lower access to inter-bank lending forces banks to reduce liquidity risk by shifting assets to be more liquid. The effect is similar to Effect 2. Thus, seen from the NFC’s point of view, it should result in a larger value drop in bank loans than securities. Secondly as bank loans are the least liquid ones the banks are expected to reduce its stock of them further.

It is important to notice that fewer security and bank-loan investments first are reflected in the market prices of securities as these are easily sold. Illiquid bank loans are reduced more slowly, as reduction more happens through limiting new loans and it is more drastic measure to redeem loans than to sell securities.

(ii) Besides the fact that higher risk requires a more liquid asset allocation, it also reduces the banks’ overall access to capital. Every time one bank lends to another the assets and the liabilities of the borrower increases and vice versa. Therefore as short-term lending is reduced and liabilities are repaid, the total asset value for banks decreases.

Consequently, the bank is forced to reduce all types of assets, not only transform them into something more liquid. Again, the effect is identical to Effect 2, thus enhancing the overall strength of the effects.

Effect 4. If depositors behave as predicted in Effect 2, money could also shift to low risk investments such as treasury bills or government bonds. Such behaviour could have the following effect:

The price on treasury bills would increase thus reducing the real interest rate premium.. As the yield on treasury or government bonds are used as a benchmark for return on other assets, the price increase will also increase prices on NFC securities as the required interest rate falls.

In reality such effect is probably limited, especially for equity valuation. The yield on NFCs security is mostly linked to systematic risk associated with the specific company and not the underlying yield on risk free assets such as government bonds. Therefore, the volatility in yields mostly relates to company risk and the effect is predicted to be limited.

Effect 5. Banks all over the world invested in American Asset Backed Securities before the crisis (IMF WEO, 2009). Defaults in the underlying assets of Asset Backed Securities and value reduction in the remaining securities resulted in:

(i) Reduced liquidity, as the overall asset value fell. The fall was twofold:

29 a. First, the value of defaulted securities was limited to whatever the debtor had left

after liquidating its assets.

b. Secondly, the value of securities that did not default was reduced, because the perceived default risk increased. Given a fixed portfolio strategy, this reduction in assets especially affected new bank loans. This should be directly observable in net-transactions.

(ii) As a consequence of reduced liquidity and lower asset value, banks’ solvency was reduced. Thus the response to lower solvency does not appear in the first period of the crisis, but only after bank assets have experienced their first decrease in value. Low solvency is expected to motivate banks to:

a. Sell off risky assets to fulfil the Basel requirement, which is the McDonough solvency ratio ≥ 8 % (Basel Committee, 2005). Liquidity is enhanced when assets are sold in favour of solid money. Securities are more liquid and therefore more volatile than loans. This speaks in favour of selling these instead of loans for two reasons: First, liquidity normally reduces the risk that assets are sold with too large discount (if we for a second ignore irrational behaviour which is also a consequence of liquid markets under specific circumstances). Loans can per definition not be sold unless they are securitized, and doing so under the crisis could result in huge discounts. Alternative they could be redeemed, but also this can result is large losses. Secondly, removing very volatile assets can improve the McDonough solvency ratio. Thus, securities are sold in favour of bank loans. Of securities, equity is expected to be sold rather than bonds, because it is more volatile.

b. For the assets which are not sold or redeemed and stored in reserves or T-bills, the banks are expected to shift strategy and:

o invest in more liquid assets (meeting the Basel requirements)

o invest in assets with higher risk of default (more risk is preferred as the solvency degree decreases).

The net-effect of lower liquidity and reduced bank solvency is expected to impact on NFCs level of finance and new finance in the following ways: NFCs’ overall level of external finance will decrease: Partly because the market value decreases (due to default and high risk premiums) and partly because banks sell securities, redeem loans and reduce new investments and lending to increase their cash holdings. For the remaining assets (that are not placed in

30 cash), high risk liquid assets are preferred by the banks. All liabilities have become more risky since also NFCs’ solvency degree has decreased. Yet, especially equity and then bonds have increased their risk premium (see Effect 8 for elaboration). Securities are more liquid than bank loans, so banks are expected to increase their holdings of securities on expense of bank loans. The effect for the NFCs is a rebound in equity and bond prices, while bank loans are expected to decrease steadily (as some are redeemed and new financing is rejected). The change security financing for NFCs, also depends on the behaviour of other security holders’

and their share of the securities market. Banks do not set the market price by themselves.

Effect 6. Banks granting loans to NFCs traditionally create a strong relationship with these.

The loans granted are usually significant to the bank as well as the NFC in order to reduce marginal cost (for both), moral hazard (for the bank) and the interest rate (for the NFC).

Enhanced short-term volatility in the security market can - under a financial crisis - reflect speculations and result in valuation not reflecting true value. A rational change in expectations to either cash flow or risk results in a larger value reductions for high risk portfolios than low risk portfolios. Large value reductions will (for reasons explained later) enhance the probability of irrational volatility in prices. Thus high risk portfolios will be more exposed to short-term irrational volatility stemming from signalling.

Banks normally do not use signalling to gather information on bank-loan debtors.

Instead they rely on screening and monitoring. This method enables them to focus on rationally assessed risk without being affected by irrational signalling and herd behaviour. It is therefore likely that banks move towards their competitive advantage; that is transforming deposits to loans. In addition the interdependence between the banks and the NFCs tends to be stronger than between NFCs and other security holders. This reduces the NFCs incentive to conduct moral hazard and increases the banks incentive to grant them the loans needed and thus reduce the default risk of the NFCs. The effect for the NFCs is (i) relatively stable flow of new loans and (ii) value reduction in securities as banks sell these at the expense of loans.