Saturday, February 02, 2013
13) The UK will NOT adopt either a Nominal GDP level or growth target.
Since George Osborne surprised the Establishment by appointing the Bank of Canada’s Mark Carney as Bank of England Governor and his subsequent comments on the potential virtues of Nominal GDP targeting, there has been much discussion and anticipation that the UK is about to abandon the inflation targeting regime and opt instead for a Nominal GDP target. Without beating about the bush, TMM think this is a load of codswallop. Notwithstanding the fact that Mr Osborne has publicly stated that any new regime would need to have demonstrable benefits over the existing one – and was reported to have had “a quiet word” with Mr Carney at Davos – TMM reckon that there are several reasons why adopting a Nominal GDP target in the UK in particular just wouldn’t work. As regular readers will attest, TMM are not fans of Mervyn King, but find themselves completely agreeing with his recent comments on the topic.
Firstly, it’s worth noting that in the late-1970s/early-1980s (at least, according to one of TMM’s colleagues who was there at the time) the Treasury had a look the possibility of such a regime, when deciding upon monetary frameworks. This was rejected as the data was not available on a timely enough basis, while also not exactly being a million miles away from targeting broad money growth & excess liquidity (i.e. how much more liquidity there is than required to support nominal GDP). We all know that this fell down as monetary velocity proved unstable.
There are here, perhaps, parallels for a potential Nominal GDP target. If real GDP growth is a function of labour force & productivity growth, setting a Nominal GDP target of say 5%, which TMM presume means inflation of 2-2.5%% and real trend growth of say 2.5-2.75%, makes two pretty big assumptions (that may or may not be correct) in addition to those of the New-Keynesian Phillips Curve. The UK has seen large changes in the size & composition of its labour force over the past decade and demographics are likely to further alter this in the coming years and decades. And while, over the very long term (decades), it seems that productivity growth has managed around a 1-1.5% trend, the Japanese experience, recent discussions in the Economist (and elsewhere) about innovation, as well as the UK’s own post-crisis productivity puzzle argue that we should not take the past trends for granted. [As an aside, TMM are optimists on the productivity & trend growth front, but there is enough evidence to suggest that we may be wrong]. The Philips Curve may prove vertical, productivity growth low, and monetary easing merely result in higher inflation as expectations lose their anchor and second round effects send wages spiraling. Inflation targeting alone makes fewer assumptions about the economy and is less complex in its operation – the 1970s experience with NAIRU demonstrates that it is difficult enough looking at just one variable without adding additional ones to muddy the waters.
Governor-elect Carney has mooted a Nominal GDP Level target, rather than a growth target, and looking at the lost output to past trends, Nominal GDP is something like 15-20% below those. TMM are highly sceptical that regaining this trend by exceptionally aggressive monetary policy will work. The gap is just too large to be credibly closed in a relatively short period of time (e.g. 5years), and the spector of the BoE buying everything under the Sun to get there provides a huge risk to inflation expectations. This of course has a corollary for the future when presumably the economy is eventually at trend and suddenly faced with a supply shock. The current inflation targeting regime allows the Bank of England to look through short term price shocks that are “one-offs”. A Nominal GDP Level target would presumably required the BoE to tighten policy when faced with such events. It is clear that the path dependency embedded in such a regime is attractive when the economy is operating below capacity (as is now), but let’s be honest with ourselves that this will not always be the case. In contrast to the implied counter-cyclical price assumption in a level target (when the level of GDP is above trend, tightening is implied, and vice versa), public inflation expectations are inherently backward-looking in their formation.
TMM’s simple conclusion to the above is that level targeting is just unrealistic, and very unlikely to get past Sir Humphrey.
So perhaps Nominal GDP growth targets are possibly a better idea?
Well, as TMM pointed out above, while more realistic in terms of policy slippage vs. gaps, these still involve trying to distill several variables into one number. And that may well work (though TMM are sceptical). But the real bug bear TMM have with this is that UK policymakers don’t know within a reasonable margin of error where the economy is *now*, primarily a result of the ONS’s inability to measure activity. As the below chart of the first release of real GDP vs. the current revision shows, the differences are huge, and not really known to any appreciable degree until 2-3 years after the event when the Income accounts have been properly reconciled from tax receipts. TMM will cheekily suggest that this represents the public sector’s misallocation of resources to Her Majesty’s Customs & Revenue, rather than to the ONS. But on a more serious note, economic policymaking is hard enough without having little confidence in the data.
The primary (though not only) reason for these large revisions is the GDP deflator. So the data is subject to large revisions, but it is also only available on a quarterly basis anyway. TMM would argue that is inferior to the timely, monthly CPI prints which are rarely revised, and significantly easier to collect & measure than activity. Of course, there will always be arguments about index construction (e.g. chain weighting, how to deal with housing/shelter costs etc), but in the grand scheme of things, these are irrelevant.
The below chart shows TMM’s attempt to get initial & final revised Nominal GDP data, which having abandoned the horrific ONS website we have settled for the intuitively simple proxy of Real GDP growth plus CPI. While the weights & constituents in CPI vs. the GDP deflator are different, TMM don’t think this invalidates the broad conclusion of this essentially broad strokes piece. Handily, this means that we can pull in the BoE’s implied Nominal GDP forecast 2-years ahead (taking real GDP & RPI/CPI wedge-adjusted forecasts assuming unchanged policy rates over the forecast period).
Proponents of Nominal GDP targeting often posit that it would have led to tighter policy in the 2003-7 period. Looking at the early GDP releases this would indeed have been the case, but the latest revisions would suggest that under the Nominal GDP targeting framework, policy would have been too tight in hindsight, as the overall nominal strength was revised away. In fact, the synthetic BoE 2yr ahead Nominal GDP average forecast for the period was 4.7%, the first print was 5.95%, but was subsequently revised to 4.9% - pretty close to the BoE’s average forecast. TMM would argue that the fact that these initially reported numbers would've likely resulted in tighter policy at that period and presumably would've prevented (or at least reduced) the credit bubble in this period is merely a coincidence caused, primarily, by the incompetence of the ONS, and should NOT be used as evidence in favour of Nominal GDP targeting. In fact, if statistical collection were better in the UK and the initial prints closer to the currently revised numbers, it is eminently possible that under this framework, policy would, in fact, have been *even looser* in the period 2004-7! [TMM will caveat that with the usual disclaimer around counter-factuals]
Interestingly enough it appears, coincidently, that in attempting to target inflation, the BoE has also effectively synthetically shadowed a Nominal GDP target of around 4.7%. This is hardly surprising, given that over the period, trend real GDP has been around 2.5-2.75% and the inflation target has been 2%. Given this, does a Nominal GDP target really provide additional flexibility for targeting growth? An ex-MPC member once admitted to TMM (though this was pre-crisis) that the bulk of what they did was to look at the PMIs and adjust policy (if needed) on the back of that (i.e. - significant attention is paid to growth, in practise). Together with the fact that until only a few months ago the BoE had the easiest monetary policy within the G4, it is hardly credible to suggest that the current inflation targeting regime does not allow plenty of flexibility for supporting growth.
Finally, and arguably most importantly, Nominal GDP growth & level targeting is far too esoteric and complex for the public to easily understand. If expectations are important (and TMM would argue they are very much so), then unpredictable outcomes may result as households & businesses try and interpret policy. For all their failings, the economic policy regimes that the UK has adopted over the decades have *all* been easily communicated to the public: the fixed exchange rate systems of Bretton Woods & ERM require no further explanation, nor do the attempts at targeting “full employment”, managing "incomes policies" by government stimuli/wage freezes etc. Even though targeting the growth rate of monetary aggregates is a somewhat complex topic, it was easily communicated and widely understood as “preventing too much money chasing too few goods and pushing up prices”. Again, targeting an inflation rate of 2% is a very simple & well-understood concept that has significantly contributed to the de-indexation of wages in both the private & public sectors in recent decades.
To conclude, TMM reckon that there is plenty to lose and not really very much to gain in departing from the current inflation targeting framework that has served the UK for the past 20 years. As noted above, it has proved amply flexible in allowing the BoE to look through supply shocks, concentrate on growth while capacity is large & wages are well-behaved. Inflation expectations have incredibly stayed well-anchored, despite the continued above-target inflation prints in recent years. This credibility has been hard won, the framework is well-understood by both the public and the markets, and could easily be “enhanced” without endangering this (perhaps by widening the range before a letter is required to be written, lengthening out the understanding of “medium term” to be 3-4 years rather than 2 years etc).
TMM reckon that a 5% Nominal GDP target (i.e. 2% inflation plus 3% growth) would merely end up being 5% inflation and zero growth. That is not something the Treasury & BoE are likely to risk in our opinion, no matter how much the transfer fee for their new star player.
With that, we wish Mr Carney (who TMM do believe was the best person for the job) the very best of luck in his new role.