Monday, 9 March 2009

Returns on bailout investment

Via Stephanie Flanders, the IMF starts to give us a way to estimate the return on investment of recent financial bailouts from various countries. Apart from anything else, this may help to silence some of the complaints from people offended at us "giving" money to banks to "bail them out".

There are two different ways to calculate ROI: first, what is the cost and benefit to the state as an entity, and second, what is the cost and benefit to the economy as a whole.

Note that the 'cost' here does not refer to the entire cost of the crisis in lost economic output or reduced asset prices - because most of that is a sunk cost and our decisions can have no impact on it. It refers to costs specifically incurred by explicit decisions taken to rescue financial institutions.

In the case of the UK, the cost to the state (according to the IMF) of the bailouts is about 9% of GDP or £130 billion. The return will include:
  1. Whatever is gained from selling off government stakes in banks during or after the crisis
  2. Increased tax revenue compared to whatever would have been raised without the intervention
On the basis of current stockmarket valuations, the government's RBS stake is worth about £5 billion and its Lloyds stake £4 billion. Northern Rock and Bradford and Bingley are no longer quoted, so there is no direct measurement of those. Of course we'd expect a strong recovery in those prices as the crisis ends, especially as there is probably a large discount in price due to the risk of further dilution or full nationalisation - a "Darling call" to balance the "Geithner put" that is considered to be supporting US bank prices.

My estimate (unsupported by data) is that government stakes in UK banks will be worth around five times the current values by the time they are sold off - RBS and Lloyds will represent much the greatest share of this, with Northern Rock and Bradford and Bingley only £1-2 billion each (about a quarter of their peak value). So that's about £50 billion - 40% of the state's money back.

Much more important is the effect on tax revenues. The financial sector rescue will preserve some level of economic activity in the sector itself - which in better years generated up to 7% of GDP and probably 10% of tax revenue. Over the next five years this may shrink to, say, 5% of revenue but this has to be compared with 2-3% if there had been no rescue - a difference of £15 billion.

But more importantly it will make the difference between a potential 7-10% drop in GDP and the actual 3-4% drop that we will get. Over the course of the next five years the actual cumulative loss will probably be about 10% of annual GDP, while a complete financial sector collapse would have cost us more like 20% - plus secondary effects from a possible collapse in sterling. The impact on tax of this 10% difference is amplified because in a recession, tax revenues fall (and social benefits rise) much more than GDP - so it's quite likely that the difference would be upwards of £150 billion. Just think of the £500 billion of deficits the government is currently expecting to see how big the impact of a recession is.

Thus, against a total cost of about £130 billion we estimate a total return of £215 billion.

So purely on the basis of the state's finances, the rescue looks worthwhile - though these numbers are of course based on many assumptions and counterfactuals which can never be tested. The biggest question is whether the collapse of the financial sector would have resulted in such a huge impact on GDP (my estimate is 10%) and whether its rescue will result in 'saving' future growth. Some people believe that the rescue just keeps an unsustainable system alive for a few years longer, though I am not in this camp.

Now we turn to national income. After all, the government is not in this to make money for "itself" but to maximise the income of us citizens.

The cost to the whole economy of the bailouts is more complex to estimate. Of the £130 billion spent by the government, some of this would have been lost by private investors instead - so the full cost of the bailout is not a cost to the economy. It could be argued that any portion which went to reduce the losses of foreign investors is a net loss to the UK economy, but then we'd want to offset that against the corresponding rescues of UK investors in foreign assets.

On the other hand, the £50 billion raised by the government from selling stakes in banks cannot be counted as a gain for the economy either, because it will simply be a transfer from new investors to the government.

So perhaps the best measure is to simply look at:
  • Does the bailout reduce economic output in any way?
  • Does it increase economic output in any way?
My estimates above are that it does increase economic output, by a total of 10% of a year's GDP, or about £150 billion.

Potential reductions in GDP could arise from a change in incentives; if investors are no longer incentivised to put their capital into the economically best option, but instead send it to places that are more risky than is ideal, relying on future government bailouts to keep it safe.

Reductions in future growth might alternatively come from increased regulation, meaning that capital will instead be pushed towards less risky investments than the ideal. The government will therefore have to be careful to avoid over-regulating the financial sector.

If we're lucky, these two factors might balance out. But without care, they probably won't.

I'd therefore propose an equivalent to the independent central bank. A new institution which, instead of guiding interest rates to hit an inflation target, tries to guide the risk preferences of investors to hit the appropriate neutral level. It is possible to calculate the rationally optimal risk preference for the economy as a whole - based on demographics and consumption smoothing across the projected lifetime of the current and future population - but even so, this is quite a task.

Something to explore anyway. If this is achieved successfully, then the whole economy's ROI for this bailout might be huge. If not, then we could easily end up with a loss.

In any case, the future success of the economy in generating GDP growth will be a far greater factor than whatever small proportion of GDP has actually been directly spent in the bailout. So don't worry about the £150 billion that the IMF is focused on. Just worry about the £20 trillion that's going to be made or lost in your lifetime according to the outcome of the next year's economic policymaking.

No comments: