Policy making in an Elysium scenario

On Wednesday evening I was fortunate enough to attend the Annual dinner of the Society of Business Economists. The keynote speaker was the Ignazio Visco, Governor of the Bank of Italy,  who gave a thoughtful and wide-ranging speech covering the whole waterfront of economic challenges. The key points for me of the speech were as follows:

  • Inequality is high and rising (especially when considering the top 0.1%). And this is true in both income terms and wealth terms.
  • Technology is developing that threatens to automate a staggeringly large proportion of jobs (Visco reckons that forecasting 45-50% of total employment will be replaced outlined here would not be outlandish), delivering a rolling disinflationary shock to the economy. The poster image of this view is the chart below that comes from Frey & Orborne‘s work.

Frey and Osborne 1frey and osborne 2

  • Macroprudential policy can be expected to be increasingly relied on to stop monetary policy stimulus feeding through into reckless misallocation of capital/ poor lending to risky projects that might undermine the stability of the financial system.

The governor had plenty of other interesting things to say (competing theories of secular stagnation, and a host of other questions), and I would encourage people to read a version of his speech here. But I don’t think I’ve misrepresented the three bullets above as three of the main points.

Now, I’ve already put my stake in the ground about where I stand on some of this stuff, and it stands in stark contrast to the bullets presented above. (I’m more of the view that there have been waves of concern stretching back pretty much forever about the degree to which technology will deliver mass unemployment and crippling real wage cuts, but these have all so far come to nothing. Meanwhile, there is a massive demographic transition coming through in front of our eyes, the upshot of which appears to me likely to deliver labour some bargaining power.) But what if he’s right and I’m just plain wrong?

Let’s run the idea ad absurdum (yes, I know, real life has a habit of getting in the way of running ideas to the extreme conclusions). But if we did:

  1. The folks who own the robots will own the means of production (as well as the means of accountancy, the means of lawyering etc etc), and there will be no room for pontificating about labour power: labour power will be kaput. Income from the sales of robot overlord kit and robot overlord output will accrue to these members of the capitalist elite, and they shall have a fine life. But given their microscopic marginal propensities to consume (and the fact that they will have become autarkic with the help of their robots) there would be a big demand problem in the economy, and disinflation (if not deflation) is the norm.
  2. If central banks continue to shy from helicopter money, focussing instead on setting the appropriate rate of interest at the short (and perhaps through QE, the long end), the days of positive nominal, let alone real, interest rates becomes but a halcyon memory. But permanently negative real rates (as set by the economy, and implemented by the central bank) inflate further the net present value of cash flows attached to the firms that own our robot overlords (in turn owned by the ultra-rich), heightening inequality outlined in #1.
  3. Given what might be decent returns to robot overlordship you might think that the sector attracts lots of fresh capital. It does, but in the form of equity finance, given the high risks associated with technology even in the robotic sector, and the macroprudential framework that has been thoughtfully put in place to guard against financial accidents that might result from speculatively financing start-ups that challenge the big players means, on the ground, that entrepreneurs are credit-rationed for the sake of financial stability. The folks that provide this equity finance are typically drawn from #1, although periodically someone does something truly clever and special, is lionised and held up as an example of how the whole system is actually very meritocratic (further buttressed by the fact that all the professional careerists go and work in the big robot overlord firms proving that the return to education is high).

Okay, I might have stretched the whole thing a little far, but let’s continue for a moment. What, exactly, is the problem that I’ve outlined? After all:

  • The economy is larger (in volume terms) and more efficient than it would have otherwise have been (or the robots wouldn’t have been able to take over). Tick?
  • There is a dynamic form of capital allocation in play (albeit in equity finance only, but that has less financial system risks anyway) to challenge the robot overlord firms, preventing complete monopolist behaviour. Tick again?

Would policymakers sit back and let this scenario play out? What happens when this sort of plutocracy collides with democracy? Put another way, what is the likely policy response to an Elysium scenario (good Blomkamp sci-fi film, underrated in my view), in which what really matters is ‘who owns the robots?’

There are some readers who might say “open your eyes, this dystopian future has been developing for decades and the answer is obvious: plutocrats win, democratic institutions are captured, and the cultural zeitgeist is also captured into knowing that the system of the day is both economically optical and morally sound“. I would counter that while it is hard to ever step out of one’s cultural context, a whole host of measures of well-being (life expectancy, health, literacy, etc) suggest that the fruits of economic growth have to a certain extent been widely-shared, and we are nowhere near the Elysium point at a national level (although this is certainly less true at a global level).

Through the instruments of the state, we collectively have in place a structure that facilitates/ optimises capitalism (contract law, property rights, state monopoly of violence etc) so that the economy can – in aggregate – grow to its potential, and we can then deal ex post with questions of distribution. This is what I understand Lord Mandelson meant when he said he was “intensely relaxed with people getting filthy rich as long as they pay their taxes“. Sort of Rawls, if perhaps not exactly. 

It seems clear to me that addressing the dim and distant Elysium scenario with a traditional monetary policy response would be problematic (see #2 above). And given that it is a question of ‘who owns the robots’ in this dim and distant scenario, I wonder whether either (a) a wealth tax, (b) a heli-money financed acquisition of robot overlord property rights, or maybe (c) a Guardian campaign to nationalise the robots would gain popularity. Disregarding option (c), I think that option (a) would be fair, but it seems that few others agree. But option (b) rhymes with my understanding as to how assignats came into existence in France in 1789: as a way to finance the acquisition of venal offices off their pre-Revolutionary holders.* (Assignats have ultimately been associated with monetary instability and hyperinflation, although it is not clear to me that this experience serves to support the quantity theory of money as much as a view that money is a microtechnology of daily trust that is ultimately somewhat fragile.) It would be strange if we found ourselves revisiting techniques used in the late eighteenth century to deal with a twenty first or twenty second century problem.

After I wrote a draft of this blog I saw on Twitter that Andy Haldane spoke this evening on all of these issues in a reasonable amount of depth. Reading the speech I think his views are very interesting, and are of course laid out in a fuller, more eloquent, and well-contextualised manner. (To be fair, he probably spent more than two tube journeys typing them up on an iPhone.) He indicates that he, like Visco, are taking the potential implications of automation seriously as a prospective intermediate disinflationary force. He ends with the idea that there are three public policy solutions that might be put into motion (relax, retrain and redistribute), and this sounds pretty reasonable (actually, Visco did make a big play for retraining now I think about it). I would have loved to have asked Haldane a question on the prospective use of helicopter money for the Elysium scenario. I hope Duncan asked it.

*Rebecca Sprang outlines in her brilliant book that ‘no one (not even Marat or Robespierre) took the truly revolutionary position of suggesting venal offices might be illegitimate privileges that could be cancelled without payment’ (p87). It seems even in Revolutionary France that radical ideas of wealth tax were considered beyond the Pale.


Inside and Outside money at the end of the yield curve: a footnote

Back in April I gave a presentation at a conference in Oxford in which I discussed some brilliant work by Branko Milanovic, Charles Goodhart and Ben Broadbent , and posited that if squinted your eyes just so, lots of things made sense in a way that might be a little against the consensus. Specifically, the asset boom story of the past thirty five years, decline of labour power and increase in inequality in the West, general disinflation and a lower trend nominal GDP in developed markets might all just be an expression of the radical globalisation and rise of China that has transformed our day to day lives. If so, then Summers’ Secular Stagnation thesis might prove to be a needlessly elegant paradox for central bankers to muddle over. Furthermore, with China approaching a Lewis point, this historical trend could be challenged and potentially thrown into reverse (higher labour power, rates, headwinds for asset prices, and some portfolio construction stuff too).

It went down sufficiently well that I wrote it up, and VoxEU kindly published it on their site giving it a wider audience than it would otherwise have had.
It has been brought to my attention that the terminology I used in the piece is perhaps either a bit confusing or confused (maybe both?). Specifically, I might be too liberal with my use of the term ‘Wicksellian rate’ for economists’ liking. This is carelessness or ignorance on my part rather than an attempt to execute an intellectual sleight of hand. So please take this blog as an extended footnote to the voxeu piece that seeks to explain:

a) what I mean by Wicksellian rate, and;

b) how changes in market expectations of this rate impact asset prices. 

It’s a bit long-winded, but I am trying to be relatively precise, which hopefully will reveal most accurately my confusion in language or content. Feedback welcome!

  1. I understand the Wicksellian rate to be the real rate of interest that keeps an economy that is running at full capacity in balance (eg, the policy rate is neither stimulative nor constrictive). This is another term for the ‘neutral rate’, as outlined by Greenspan here, Broadbent here, and Bernanke here. (I am aware that there is a big literature on whether the neutral real rate is even a legitimate notion to entertain, but have so far through my meagre readings been insufficiently persuaded to drop the notion, and enjoyed Brad DeLong’s recent post on the matter.)
  2. We live in a monetary economy that features outside money (central bank/ state money), and inside money (private money invented by banks in the form of deposits every time a loan is made, and to a meaningful extent invented by anyone who considers an asset to have moneyness and can persuade others of this feature). The Wicksellian rate will be a rate that consists of some mixture of the expected or inferred rate of return on some mixture of outside and inside monies. It would be nice if central banks told us the mix they thought important as this would help market folk like me understand their reaction functions better. And from time to time we do get hints (like footnote 24 in Carney’s Jackson Hole speech, the prominence or otherwise of mortgage spread charts in Bank of England QIRs, etc). And maybe that is fair enough given that central bankers are just doing their best to guesstimate through the fog.
  3. But the Wicksellian rate is set by the economy and not by the central bank. Central Banks seek to guesstimate where it might roughly be (including things like credit premia, liquidity premia etc), and they then manage their nominal (outside money) policy rate and expectations thereof around this guesstimated real rate to either stimulate or restrict demand and return the economy to it’s long run potential.
  4. Monetary sovereign bond markets (eg, the Gilt market in the UK, the US Treasury bond market in the US) can be thought of from a bunch of angles, one of which is that they consist of market participants’ best guesses as to where this overnight rate might be over time (ok, plus a guesstimated term premia, guesstimated liquidity premia, guesstimated premia associated with a raft of regulatory provisions on banks etc). And so by looking at the yields on Gilts or US Treasuries with different maturities you can back-out market participants’ expected term-structure of rate expectations set by the Bank if England or the Federal Reserve respectively.
  5. These monetary sovereign bond yield curves can furthermore be decomposed into a guesstimated path of real rates and a guesstimated path of inflation. And the real part of the curve (eg, inflation linked real rate curve) can be thought of as the market’s best guess at the term structure of the (outside money) Wicksellian rate, +/- near term cycle stuff.
  6. So actually, the long(ish) forward real rate is going to be the market’s best guess as to the (outside money) Wicksellian rate for the economy.
  7. Inside money yield curves are priced off outside money curves. The starting point when pricing term credit (let’s say, the interest rate that a company might have to pay to borrow on a fixed rate basis when borrowing for ten years) is to observe the monetary sovereign yield for that term, and add additional yield premia to compensate for estimated higher liquidity costs, estimated credit costs, some premium to reimburse the costs associated with performing credit analysis etc. (Fellow bond geeks may take issue with this point as being too massive a simplification, but I don’t think much is lost leaving this description in place.)
  8. So you get lots and lots and lots of private (inside money) yield curves describing the market yields associated with different term borrowings of different private borrowers. Many if not most will feature only one point.
  9. These curves consist of yields (aka discount rates) that are used to discount expected cash flows and calculate the asset price in a discounted cash flow (DCF) analysis. If the cash flows are unchanging, a falling yield equates to a rising price and vice-versa.
  10. We can also use DCF when looking at non-bond asset pricing. It’s just that we have to imagine what the cash flows might be and also imagine what risk premia to apply.
  11. So after you have finished building a model of company earnings, you still need to imagine into existence an appropriate discount rate to apply to your model’s resultant cash flows to calculate your assessment of its ‘fair value’. Given that equity is the most junior liability in a company’s capital structure, the discount factor that you would apply to these equity cash flows might still be higher than the one applied to the same company’s most junior debt obligation, which in turn would be higher than its more senior obligations of equivalent term. The observable (inside money) discount factor/ yield curve would in turn be priced as higher than the monetary sovereign’s yield of equivalent term (eg, point 7 above).
  12. I take a bunch of approaches when I look at equity discount rates. A super-simple one is to look at the trading relationship between corporate debt and consensus forward earnings yields. One is a nominal yield, the other one is generally believed to be a real one. But the nominal one contains an estimate of real rates and an estimate of inflation.
  13. With only marginally more complexity, I solve the equity internal rate of return from current equity prices, consensus earnings projections for as far as they are available, and estimates for residual nominal earnings growth rates taken from historical relationships between nominal earnings and GDP, assumptions about profit share of GDP and payout ratios. CAPM is another approach, and I think this one is interesting too.
  14. To cut a long (enough) story a little short, underlying all private discount rates sits the outside money (government bond) yield curve with its pricing of anticipated inflation, near term policy developments, and long-term guesstimates of the (outside money) Wicksellian rate. Changes to that last bit change asset prices all over the shop.

So while it is clear to Woodford et al that the Wicksellian rate is a real variable relating to the marginal return on capital rather than a real (eg, inflation-linked) market rate, in my analysis it is also the market’s best guess of the likely medium to long-run unobserveable equilibrium (Wicksellian) real interest rate. And crucially, as the market expectation of this real rate falls, so asset prices (the present value of cash flows that use this falling market rate in some way, eg, #7-10 above) rise correspondingly.

This raises what now feels like a perennial market-circularity question: to what extent have rising asset prices been a function of diminishing rather than improving prospects in the West? (Eg, when weaker cash flows are accompanied by a lower discount rate for those cash flows, the asset price implication is ambiguous rather than negative.) Or, in the phraseology of my voxeu blog, to what extent might demographically-driven resurgence in labour power raise the Wicksellian rate in the West, and with it deliver a headwind for asset prices and challenges for portfolio construction?

In my day job I have plenty to say on this, but it would be inappropriate to say anything on my private blog. Instead, here’s a link of me and Joe Weisenthal chatting about it a couple of weeks back.