FISIM

Larry Elliott brings us an interesting story:

the answer to the puzzle lies in the way the ONS measures the value of financial services. Although this has a name that only a number-cruncher could come up with – Financial Intermediation Services Indirectly Measured (FISIM) – it is actually quite a simple concept. Banks charge interest on the loans they make and they pay interest on the deposits they take in. FISIM is calculated by subtracting the interest rate on deposits from the interest rate on loans and multiplying by the number of outstanding bank balances.

What happened in the fourth quarter of 2008 was that banks assumed there would be a massive increase in defaults on loans. They responded, entirely rationally, by increasing interest rates to cover the expected losses. That meant the gap (spread) between interest rates on deposits and interest rates on loans widened, the value of the financial sector\’s services as measured by FISIM increased, and this showed up in the national accounts as an increase in output.

\”In other words,\” Haldane says, \”at times when risk is rising, the contribution of the financial sector to the real economy may be overestimated.\” If he is right, the recession during the winter of 2008-09 was probably even worse than the official statistics suggest.

OK, yes, we all know GDP ain\’t everything and nor is the way it\’s measured. But this FISIM is an attempt to measure the value added by the financial or banking system so that we can shoehorn that added value into our measure of value added across the economy, GDP.

What Larry is arguing is that increased interest margins aren\’t in fact added value and thus we\’re over-measuring the value added by banking.

However, rising interest margins are added value.

Start from the beginning. Yes, we want a financial system. We want a payments method, we want a method of maturity transformation. We\’d also like someone to stand in the middle with their capital carrying the risk of default.

So the existence of a financial/banking system has value. In a market system that value is going to be measured by what people are prepared to pay for those services. When risk is, as noted, rising, what people are willing to pay for those services will rise. This is, of course, exactly the justification for the bail out of the system: it\’s far too valuable to be allowed to collapse into smouldering ruins (even if certain institutions within it should be happily left to go bust).

Thus the very fact that people are willing to pay more is an indication that people value the value added more….and thus that, when risk is rising, the value of banking services rises and this should rightly be reflected in GDP.

Don\’t forget, GDP is the *value* of goods and services produced and we can have changes in GDP without there being any change in quantity of either goods or services produced. A change in the value we, consumers, put upon them is a change in GDP just as much as a change in quantity is.

Put it another, simpler way. Imagine that all a bank did was provide a lock box for your cash. Pretend (yes, pretend, this is a model, a thought experiment, with all of the ignoring tiresome details that that entails) that this lock box was 100% secure.

So, in a society with no crime little value will be associated with either the existence or use of these lock boxes. Now let\’s change the risk of theft. The land is covered with feuding, feudal, tribes, all trying to nick your cash. Those lock boxes are still 100% secure. Risk in this economy has definitely increased. So, also, has the value that we would put upon the existence and or use of those lock boxes: of the services that banks provide.

Even while the size of the economy is being devastated by the depredations of those feudal tribes (The Anglo Taxons, etc) the value we ascribe to banking will be rising and thus banking\’s contribution to GDP will be rising.

Agreed, it\’s a pretty odd model but it doesn\’t look like, prima facie, that an increase in risk cannot be associated with an increase in the contribution of banking to GDP.

4 thoughts on “FISIM”

  1. Not my field of expertise, but aren’t there two different risks here – the customer’s risk (which you’re looking at) and the bank’s risk of customer default (that Elliott is talking about)?

    And isn’t the price of the banks’ intermediation services actually:
    (interest charged) minus (interest paid) minus (bad debts written off by the bank) ?

    Of course alternatively we could just take the full value of the interest charged, and say that that’s the value of the bank’s services (as valued by its customers), but that isn’t what GDP does. Instead it tries to put a price on the pure intermediation of the bank sitting in between the savers and the borrowers. In which case it seems legitimate (as Elliott’s argument seems to run) to adjust for bad debts.

  2. So, in a society with no crime little value will be associated with either the existence or use of these lock boxes

    I disagree. Money stuffed into my mattress at home may be safe from thieves in this zero-crime society, but:
    * it is at risk from fire, flooding etc
    * I can’t easily access it from a cash machine when away from home.

    I would like to see current accounts become these virtual lock boxes. This would be the only deposits guaranteed by the government, precisely because the money can never be at risk (in terms of lending it out).

    Note that National Savings doesn’t count, as the government spends the money savers deposit, it doesn’t just store it.

    Government also needs to guarantee the technology infrastructure that runs the ATMs etc. The fact that many banks have outsourced their data centres to IBM, EDS etc may make that easier.

    Let the rest fail if it chooses to.

    Even while the size of the economy is being devastated by the depredations of those feudal tribes (The Anglo Taxons, etc)

    I’m not sure if this is a typo, or if you are describing HMRC as a feudal tribe 🙂

Leave a Reply

Your email address will not be published. Required fields are marked *