The Risky Lending Gap

 This blogpost steps back from the details of the NAMA-nationalisation controversies and focuses on fundamental concepts.  What policy goals are we trying to achieve and why are these goals important? 

The key policy goal is to close what I call the risky lending gap – defined as the difference between the lending policy of a sound bank and that of a bank encumbered with an overhang of poor-quality loans.  Filling this risky lending gap (if it exists) is important for jobs, growth and exports.  Possible policy tools (as Patrick Honohan has made clear in earlier posts and his recent presentation) include a capital injection, or an asset switch, or a combination. 

I will rely on a technical model whose details are given here.  First I consider the case of “good bank” with equity capital Q which has to choose an optimal lending policy.  The bank is faced with a supply of available lending opportunities, ranging from riskless up through increasing degrees of riskiness.  The “safe” loans (which are the first loans that the bank will include in its loan book) might be mortgages to senior civil servants, the riskier loans might be loans to new entrepreneurs planning to hire recent graduates and open export businesses.  As policymakers, we are concerned particularly about the more-risky (but economically valuable) commercial loans, since these are the source of new jobs, positive growth, and increased exports.  For simplicity I assume that the riskiness of the bank’s loan book increases linearly in the amount of lending:

 

Making some technical assumptions (constant risk aversion, normally distributed returns on the loan book) one can derive the unimpaired bank’s total lending.  I assume that this is the socially-optimal amount of lending, since the bank and loan markets are functioning properly in this case:

 

The model is calibrated to roughly approximate a medium-large Irish commercial bank such as Bank of Ireland or Allied Irish Bank. See the spreadsheet here for details of the calculations. The bank has equity capital of 10 billion Euros and optimal lending by the bank is 90 billion Euros. 

Now I consider the case of an impaired bank. The bank “inherits” a stock of bad loans from earlier bad decisions and/or bad luck (a combination of these, obviously, for the case of Irish banks circa 2009).  These developer loan losses D are an overhang of net losses on the bank and are unaffected by the bank’s choice of current lending.  They are also risky in terms of their realized value, and this risk is correlated with the risk of the bank’s new lending.  The next graph shows the bank-optimal lending level given this overhang of bad loans, and compares it to the socially-optimal bank lending of 90 billion derived in the earlier graph.   Because of higher risk from the overhang of bad loans, the bank cuts back sharply on new lending.  Bank lending fall from 90 billion to 73 billion Euros – this is the risky lending gap.  Since small-business loans are riskier than mortgages, this represents 17 billion Euros of lost new business lending which otherwise would create jobs, spur GDP growth and increase exports.  This is the problem that NAMA, the Irish government’s bank capital injections, the Geithner plan in the USA, etc., seek to address.  

 

 

What policy ideas does this simply model clarify?  Here are three:

1.  How big is the risky lending gap?

New Irish bank lending has gone down since 2007, but this is not necessarily due to a risky lending gap.  The supply of lending opportunities shifts unfavourably during a recession, so even socially-optimal lending will fall.  Patrick Honohan spoke about this in his recent DEW presentation.  The research problem requires more data on actual loan activity and more detailed empirical analysis, but is important to measure the risky lending gap in the current Irish recession.  This is a worthwhile endeavour for government, academic or practitioner economists, given that the stakes are so high.  Perhaps the true risky lending gap is near zero, and NAMA is a waste of time and money. 

2.  Should Policymakers Focus on an Asset Switch or a Capital Injection?

The source of the risky lending gap could be the expected size of the developer loan losses or it could be their uncertainty.   The appropriate policy response differs in these two cases.  If it is the absolute size of the expected losses, then the bank needs a capital injection, and there is no particular reason to have an asset swap.  On the other hand, if it is the riskiness of the realized loan losses, then an asset switch (where NAMA takes bad loans in exchange for a fee or financial claim on bank assets) is the appropriate response.  If both the size and uncertainty of the loan losses play a role, then perhaps a combination policy (asset switch plus capital injection) is appropriate.

3. Commercial Banks Are Information-Relationship Businesses Not Asset Portfolios

One important caveat is that, for the purposes of policy analysis, banks must be treated as information-and-relationship businesses, not simple asset portfolios.  An asset switch removing D from the bank’s balance sheet and replacing it with some other liability is not like buying one share and selling another in a stock portfolio.  The realized value of D can depend upon who is managing the impaired loans.  Also, it is very important to avoid political interference in loan recovery management.  If bank nationalization or NAMA might cause a big increase in the realized magnitude of developer loan losses then perhaps a do-nothing policy is better.  Irish property developers are very politically powerful, due to their long history of generous political donations.  When considering government policy and its effectiveness, it is important to keep in mind that Ireland is not Scandinavia.  We may have a sprinkling of their fair hair and pale skin among the populace, but when it comes to standards of governance we are a bit more, um, carefree than the Scandinavians.  We have a long and sorry history of political corruption associated with property development.  This argues for a minimal-intervention policy when dealing with bank capital and property development, where the temptations are enormous for resource-siphoning to special interests.

References

Honohan, Patrick (2009) Presentation to the Dublin Economic Workshop, May 20^th 2009.  Presentation slides available at www.irisheconomy.ie.