The wrongest thing ever? My defence of multi-billionaire hedge fund inflation truthers

I recently bought a copy of the 1958 Radcliffe Report, prompting Lorcan Roche Kelly to ask whether I was ‘trying to find the wrongest thing ever?’. I have yet to read this 350 page report that I so enthusiastically purchased, but the phrase sprang to mind when thinking about what to call this defence of a hedge fund billionaire inflation truther that I have somewhat weirdly written.

QE was expected by many to result in higher levels of inflation. It is generally understood by most researchers that QE boosted inflation versus the counterfactual.

Some people – generically speaking, those who believe the quantity theory of money is a reasonable way to understand the world (and understand that: a. money to equal central bank reserves, and also that; b. the velocity of said money should be stable over the medium term rather than just be a flexible residual in a pretty silly identity) – feared that massive levels of inflation – even hyperinflation – would result from QE. Despite these fears, inflation has been so far conspicuous mainly by its absence.

Paul Singer had a bit of a rant recently arguing essentially that QE is delivering high levels of inflation that is best observed in contemporary art prices and Hampton real estate. Sadly, I am going to have to insert a link to ZeroHedge for people who might want to read the original rant. Paul Singer is a smart guy, but I’d be more inclined to agree with Alex Turnbull’s Twitterstorm takedown:
Singer’s client letter taps into the spirit of CPI truthers: a large and disparate group who believe that the government is manipulating reported inflation numbers to hide the truth of spiralling prices that any thinking person can see if they look.

The truth is that CPI truthers are generally nuts. Their number and the persistence of their message speaks to the collapse in trust in the most basic government functions (which is appropriate in some EMs, and most autocratic regimes). If you think that ONS is doing something poorly thought out, go and have a flick through the ONS’s CPI Technical Manual before kicking up a fuss. Besides anything else, it is a really surprisingly interesting and thoughtful document (with discussions of things like plutocratic versus democratic weighting methodologies).

Despite all this, I have some sympathy with Singer and a qualified sympathy with the underlying sentiment of the truthers. My sympathy is, however, based on an idea of mine that I recognise as somewhat more fruity than Singer’s: the under-reporting of pension saving costs in the CPI.

If there was one thing that most people who think for more than a second about the CPI would probably agree on, it is that savings have absolutely no place in a consumption index. Too stoopid even for me you might think.

If we think of pension contributions as regular savings into a piggy bank that is smashed upon retirement, with whatever happens to be in there then spent, I would agree that pension contributions do indeed have no place in any CPI. But I don’t think that this is really what pension contributions are. Instead, when you put money into a pension (I’m thinking really of making a Defined Contribution Scheme contribution in this instance), you are purchasing a series of claims on other economic agents (typically through company profits that will need to be siphoned off to meet your debt, rental or dividend payments, or – in the case of government bond purchases –taxes on the general population to fund your debt interest, absent seigniorage).

These various claims (or assets as most people call them) will have estimated internal rates of return (or discount rates). For bonds these would best approximate redemption yields to worst; for equities these would be dependent on your equity risk premia model, etc. Changes in discount rates would make a pension payment that you are aiming to deliver post-retirement cheaper (if discount rates went up) or more expensive (if discount rates went down).

This change in the cost in purchasing your future pension is completely absent from the CPI. And, as a reminder, if you are looking for a retirement income of say £20k per annum, you should – according to this rather out of date but easily Googleable Guardian article, be spending about a quarter of your income on the purchase of claims that can be accessed in retirement (eg pension contributions). So this is not a little missing component.

To simplify, let’s imagine all pension contributions go to buy claims that looks like a 20yr zero coupon government bond. In making a DC pension contribution of £100 in June 2007 you would have been buying a future cash flow (pension payment) due June 2027 of £259.78. In 2014, how much would it cost to buy a pension payment in 2034 of £259.78? The answer is £131.77. The annualised ‘inflation’ of this form of pension contribution was 4.0%. This is somewhat higher than the 3.1% level of inflation associated with the CPI. To be clear, this is not the return of zero coupon Gilts over the period, but rather the difference in cash contributions required to buy the same series of terminal cashflows. It should be said that I have not deflated these terminal cashflows to keep them constant in real terms, just nominal terms. The rate of inflation attached to purchasing the same real cashflows is somewhat higher, but things do get a bit more complicated (and this example is supposed to be simple).

And so despite being a bleeding heart liberal, I can sort of understand where Singer is coming from when he talks about asset price inflation mattering for reasons more than those that Turnbull is clearly right to identify (problem credit created for bubbly stuff). But for this to be true we might need to do something whacky in the way that we think about inflation. Which is probably going to be a bit of a stretch for most. It is still a bit of a stretch for me. But I can see the argument and find it hard to dismiss.


@T0nyYates wrote a thoughtful reflection on this blog that I think is worth directing people to here.