Wednesday, January 12, 2011

QE2 revisited

When I talked about QE2 before, I said the stimulus of $2 billion in 1933 was less than the $800 billion the Fed is undertaking now.  What I didn't realise is how much less.  The 1933 $2 billion  is about $35 billion in today's money, since price levels have risen 17.5 times since then.  But the real economy has also increased 17 times since 1933.  Now money/debt/bank deposits/loans etc are nominal amounts.  So any monetary stimulus (whether tax cuts, spending increases or "open market operations"/"quantitative easing") have to compared to nominal levels of economic activity, not real (inflation-adjusted) levels.  In nominal terms, economic activity in the US has increased around 300 times since 1933.  Thus to achieve the same effect of the $2 billion then, relative to economic activity, the QE stimulus would have to have been $600 billion, and it fact it's one third higher.  This stimulus is massive.

A brief aside on the quantity theory of money.  Nineteenth century economists postulated a relationship between money supply and economic activity, in a formula like this:  MV=PT.  To de-mystify: M = the money supply; V= the velocity of money (the number of times it "turns over" -- if everybody is paid weekly you'll need a smaller money supply than if everybody is paid monthly); P=the price level, and T= the level (in volume terms) of "trade", i.e., GDP or national income.  The experts will elaborate all the complexities raised by the precise definitions involved, but for the time being, forget those.  It's the principle which matters.

What the Fed is doing is massively expanding the money supply/easing credit availablity.  This implies either that velocity will collapse (because low confidence will stop ppl spending more); or that the price level will rise with no increase in T; or that T will rise.  Now a price level rise is very unlikely.  If this were Zimbabwe, with a kleptocratic hierarchy, where hyper-inflation expectations are embedded, this is exactly what would happen.  In fact V would rise, as everybody tried to dump their worthless currency even faster.  But in the US, price expectations are for very low inflation over the next few years (look at the yield on index-linked US treasuries).  In fact, a modest rise in prices in the US would be a good thing.  Only rising commodity prices are keeping inflation positive in the US, and those rising prices are due to the Chinese boom not the US one.  Underlying inflation is close to zero.

That leaves V and T.  If confidence is so catastrophically low that massive increases in liquidity (M) do not translate into increases in T (real economic output and demand), then V will fall, and the stimulus will fail.  But confidence isn't that dire.  True, it's not marvellous.  But it's enough for new hiring to take place, for some new investment in plant and equipment to occur, for retail sales to hold up.  So even if some of the new M is absorbed by lower V, some will feed through into T, alias real growth.  And as that picks-up, so some of the stored M will be spent, and V will start to rise.  In turn this will expand T which will again improve confidence leading to further rises in V and so on.  The virtuous cycle.

Of course, at some point M will have to be restrained if P isn't to start accelerating.  That's the trick.  And it's a hard one.  But it's next year's problem.  I am no longer perplexed.  I think economic growth will accelerate this year, dragging the market with it.  Next year, though, is murkier.  Withdrawal of the additional M and deficit cutting round the globe will slow growth.  But that's 2012's problem, which I'll talk about later.  And of course, the structural problem of too much debt (private and public) remains.