In the February Inflation Report we saw the Bank forecasting (and hence, targeting) CPI inflation significantly above 2% on the two year horizon for the first time in four years. In this week’s Inflation Report we saw the Bank revising up its real GDP forecasts for what Chris Giles says is the first time since 2007. What a strange co-incidence that is, eh?
Claire Jones has a nice post covering the improvements to the Inflation Report prompted by the Stockton Review. For the first time, I didn’t have to wait a week for the Bank to publish their sacred Excel spreadsheet with the forecast data. Rejoice! This graph shows how the median forecasts of the CPI rate have moved over the last four Inflation Reports:
The median forecasts have shifted down across the entire forecast period, since February, and now perfectly hit 2.0% CPI on the two year horizon (versus 2.3% in February on the same horizon). The obvious response is to castigate the MPC for yet another opportunistic disinflation. In this case I wonder whether the Bank’s models might only have altered the real/inflation split, since the real GDP forecasts have moved in the opposite direction to inflation; Chris’ post has the graph showing the latter. It would be useful to have the forecast for the path of nominal GDP so we could identify such cases.
It is worth noting here that Mervyn King’s hawkish ITV interview in March seems to have “successfully” capped the rise in market inflation expectations seen earlier in the year, and put a floor under Sterling – at least the latter of these feeds in to the Bank’s forecasting model. King declared in that interview that the pound was “close to properly valued” and insisted the Bank was not going soft on inflation. Great work, Merv!
(The sharp movement in market inflation expectations at the start of January is not a data error, it was caused by the outcome of the RPI methodology consultation.)
So the usual conclusions must be drawn… does the MPC want higher inflation? No. Are they constrained from moving inflation expectations? No. Have the MPC been desperately printing money to raise (or keep elevated!) expected growth and inflation since February? No, no, no.
Christina Romer’s paper links to this brilliant “infomercial” video from the US in 1933, which explains the power of monetary policy – which, by the way, is all about expectations:
I had been unsure what I could write about King’s legacy at the Bank of England, but the Governor provided a quote which captures it well. In the press conference today, this is how King responded when asked about the new remit, and whether the Bank needed more discretion:
No I don’t think we need more discretion as such, I think the two key features of the remit which I welcome are, one – reaffirming that the central objective of monetary policy, which is the main role of a central bank, is to meet the inflation target of 2% a year; in other words the commitment to price stability. Nothing is more important than that and the Chancellor has moved – well he wasn’t tempted to go down the path of giving up the target, he’s reaffirmed the commitment to price stability.
Mervyn King coined the term “inflation nutter” in 1997 to describe those who embraced stability of inflation above all else. Governor King in 2013, along with an entire generation of central bankers, will still claim “nothing is more important” than price stability. That is the legacy of King and the rest of the modern-day inflation nutters.
The “revisions” critique of nominal GDP targeting seems ever more absurd as time passes. The argument pushed by some is that it’s hard for monetary policy to target something which gets revised, like the nominal GDP statistics, but we can target the CPI because that does not get revised.
So here is brief a look at this year’s two new ONS series. First is the CPIH, which is a revised version of the CPI now including owner-occupier housing costs. Second is the RPIJ, which is a revised version of the RPI, switched to use the Jevons formula. The UK’s inflation target between 1993 and 2004 was a variant of the RPI which excluded mortgage payments, the RPIX series. The ONS now lack confidence in both RPI and RPIX to such a degree that the series are no longer designated as “national statistics”.
Given these revisions to the price index methodology, is there any change of heart from the inflation-targeters? We spent ten years targeting a price index which is now considered to be of a poor quality! Was that not a bad thing? Was not UK monetary policy “wrong”, ex post, because we now have a different – better – way to calculate inflation? Or doesn’t it matter?
The answer is surely that it doesn’t matter. We already dumped the 2.5% RPIX target in favour of a new target using a “better” methodology, 2% on the CPI, and I don’t recall anybody making hysterical arguments about how badly misguided UK monetary policy had been under the RPIX target. Methodology changes, and policy adapts.
Here is a graph of the cumulative “error” in the RPI and CPI methodology, compared to RPIJ and CPIH respectively:
Both the RPI and CPI “overestimated” inflation in comparison with the new indices. The total change in prices over fifteen years measured by the RPIJ was 6% smaller than the change in prices measured by the RPI. The change in prices over eight years measured by the CPIH was nearly 2% smaller than the change measured by the CPI.
The latter is particularly interesting; for all the talk of the Bank of England “blowing bubbles” by ignoring soaring housing prices, the inclusion of owner-occupier housing has pulled down the price index since 2005. Targeting such a price index would, if anything, have allowed slightly easier – not tighter monetary policy over the last eight years.
Buttonwood at the Economist, followed by Chris Dillow [edit: fixed link] and Duncan Weldon, are all concerned about weak Sterling. I’ll pick on Chris, who considers whether the fall in the pound this year will raise prices in the UK:
Perhaps, then, the weak pound won’t add much to inflation – though this’ll rise anyway in the next few months for other reasons. History tells us as much. Since 1990, there’s been little link between moves in sterling and in inflation. The devaluation in 1992 did not prevent inflation falling sharply. Nor did sterling’s falls in 2000 and 2003 or its large fall in 2008 have a big effect on inflation.
You might think all this is good news.
It’s not. For one thing, if UK importers and wholesalers can’t pass on higher import costs their profits will be squeezed; this is an especial danger for smaller firms lacking market power. Granted, they’ll try and fight against this by cutting other costs. But this could well mean cutting jobs. In this sense, sterling’s fall could raise unemployment.
It is frankly bizarre to see these right-minded people argue against devaluation for a depressed economy. Would they have made the same case in 1992 when the UK had the wrong monetary policy (Germany’s) or in the early 1930s after the UK had suffered from bad monetary policy, re-entering the gold standard at the wrong level after WW1?
Chris’ doubts about whether Sterling devaluation will add to domestic inflation neatly invokes Bernanke’s reductio ad absurdum: if we can print money, buy foreign assets, and domestic prices never increase in response, we have discovered the biggest free lunch in the history of the world. We could have the Bank print money and buy up every single asset in the rest of the world, and then live a life of luxury off the flow of income. Shall we start with Spain?
This is wrong, and devaluation always “works” if you want to raise domestic prices and spending, because devaluation is monetary phenomenon. I find it much easier to think about the world in terms of nominal GDP and money, than inflation and output (real GDP). If the Bank prints money and buys stuff, whether that is Euros or gilts, then somebody has to hold more money. It is this “excess” supply of money, which bounces around as a “hot potato”, which results in increased spending – more NGDP. Whether or not that results in rising prices in the short run depends on the supply-side response, but in extremis (the Bank trying to buy up the entire world) it must. An expected increase in the supply of money will also raise the velocity of money today flowing around the economy. If that is not intuitive then browse through the results of a Google search for the terms “devaluation panic buying” . “Panic buying” is another name for a spike in the velocity of money.
The UK’s experiment with the European Exchange Rate Mechanism between 1990 and 1992 provides a very clear demonstration of all this. John Major and Norman Lamont very deliberately tightened UK monetary policy in 1990 when they entered the ERM, pegging the pound to the Deutsche Mark at a rather high level. This “worked” by bringing down nominal GDP growth from the 8% to 10%+ annual rates since through most of the 1980s, to a 3% annual rate by late 1992:
The pound’s overvaluation under ERM went hand in hand with tight money, just as devaluation and easier money (faster NGDP growth) went together after Black Wednesday. Inflation was pretty moderate even with much faster NGDP growth after 1992, because UK economy had significant spare supply capacity. If you believe the same applies to the UK economy in 2013, you should welcome devaluation as a sign that monetary policy is being relaxed.
I enjoy looking at the historical data; the annual GDP deflator for 2012 at 1.3% is the fifth lowest on record (after 1954, 1959, 2000 and 2009). Will the hawks celebrate this impressive result?
In the picture of tight money below, fiscal fiddling with the VAT rate shows up as the large divergence between the GDP (market price) and GVA (basic price) deflators between late 2008 and 2012.
Some eyebrows have been raised after Mark Carney used the phrase “flexible inflation targeting” to describe the current UK monetary policy regime. I don’t think that should be a surprise, and I do think it’s correct to call the current regime “flexible inflation targeting”.
I keep saying it, but I think the source of confusion here is a failure to appreciate that inflation targeting is (and monetary policy generally can only be) forward-looking. Below I’ve quoted from a speech made by Mervyn King back in 1994 when he was merely the Bank of England Chief Economist, explaining this key point (after dismissing the need for an “intermediate” target for money supply growth alongside the formal inflation target):
The use of an inflation target does not mean that there is no intermediate target. Rather, the intermediate target is the expected level of inflation at some future date chosen to allow for the lag between changes in interest rates and the resulting changes in inflation. In practice, we use a forecasting horizon of two years. It is absolutely crucial to understand that the inflation target does not mean that policy is set according to the current rate of inflation. The latest inflation rate is relevant only in so far as it affects the projection for inflation some two years ahead.
My emphasis added. I can only hope that people read and re-read those last two sentences until the impact fully sinks in. To concentrate the mind, imagine that you are on the MPC in the Autumn of 2011, when the CPI rate is nearing 5%, and almost the entire press corps is berating you for printing too much money and being “too flexible” about the 2% target.
The policy regime King was following in 2011 is the same policy regime he was talking about in 1994, and that’s broadly the same regime that the Bank has followed every year since it gained independence in 1998.
King’s speech gives a standard explanation of flexible inflation targeting, or “inflation forecast-targeting”; where the monetary authority is flexible to the point of ambivalence about the current, observed rate of inflation. The Bank’s aim is to set policy such that its own forecast for inflation is at 2% at some point in the future. The Bank tends to use the two year horizon, although the MPC never specifies that precisely.
Flexibility Does Not Imply Discretion
I am rehashing an old post here, but this is where things get interesting. If you accept this interpretation of the current policy regime – be flexible about the current CPI rate, and keep the forecast of the CPI rate on target, we get a simple measure of the Bank’s policy stance: the deviation of forecast inflation from desired inflation.
To simplify a bit, I think it’s correct to say that “flexibility” in this specific sense does not involve the Bank exercising “discretion”; the choices the MPC make should be predictable and derive from the regime itself, plus their forecast model. A “neutral” policy stance should mean that Bank is always targeting 2% inflation.
If the Bank has set policy such that their forecast of inflation is not on target, they are exercising some kind of discretion, following a different policy than prescribed by the regime.
The chart below shows what the Bank’s own published data says about their policy stance over the last nine years under the 2% CPI target, showing the forecast for CPI inflation looking both two and three years forward, at each quarterly Inflation Report:
This shows the significant variation in the median forecast on the two and three year horizon in recent years. Those two downward spikes in late 2008 and late 2011 are exceptional. The Bank was not targeting (i.e. had not set policy such that the forecast equals) 2% inflation on the two or three year horizons at those times. Why not? An obvious correlation is that current observed UK CPI inflation was very high, exceeding 5% in exactly those same time periods; late 2008 and late 2011.
In my opinion this is evidence of a discretionary, disinflationary bias in the MPC. Specifically, monetary policy has been almost continually “too tight” even if you accept that “target 2% inflation on the two year horizon” is the optimal policy regime. There are two obvious counter-arguments:
a) “The Bank is impotent at the ZLB. They can’t raise inflation whatever you say about their forecasts.”
b) “The Bank may be able to affect inflation and demand at the ZLB by printing money, but their forecasting model can’t capture the effects of QE so they are unable to make the forecast dependant on their actions.”
I think (a) is obviously false, but more importantly the Bank think it is false; otherwise they would not have done QE, stopped QE, started QE, etc, etc. This is also no excuse for the November 2008 forecasts being so far off target when interest rates were around or above 3% right across the yield curve, with no zero bound in sight.
(b) might actually be true, I’m not sure. If it is true, this is even stronger evidence of discretionary behaviour; the MPC has no basis on which to choose whether to do £100bn of QE or £50bn if they have no way to model how that money affects inflation and demand outcomes.
Back to the Future
The Inflation Report will be a fun one this week, and I’m looking forward to reading the minutes from this month’s MPC meeting. The unusual statement which accompanied the MPC decision implies their own forecasts have the CPI rate back above 2% on the two year horizon. The last time this happened was Spring of 2011 when the hawks were fighting to raise rates; the wording in the statement implies perhaps that the hawks have declared a cease fire. I’m using the language of discretion, here, of course.
I have to end this rambling post with an irresistible quote from Milton Friedman in 1960 which King used in his 1994 speech:
‘The failure of government to provide a stable monetary framework has thus been a major if not the major factor accounting for our really severe inflations and depressions’.
King gave his future self a get-out-of-jail-free card, insisting that “real shocks” can also have the same effect as bad monetary policy… a position which looks rather convenient for the UK’s top central banker in 2013. I’d bet future historians will go with Friedman.
This is a post attempting to illustrate how seriously we should take the idea that there’s a thing called “the price level”, and that it’s the one thing we can measure so accurately that it should be used as the nominal anchor.
I’ve graphed the change in the level of all the broad price indices for which the ONS have a data series. (I couldn’t find a series for the old “Tax and Prices Index” which Andrew Sentance mentioned.)
Has the UK “price level” gone up 20% over the last five years or less than 10%? You tell me.
GDE = Gross Domestic Expenditure, so the deflator is for everything counted as “Consumption” or “Investment” in the national accounts. GDP is measured at market prices; GVA is the “basic price” deflator, so ignores changes in indirect taxes which do affect market prices.
I should have put this in my previous post. Martin Weale professed his concern for gilt yields:
But if we were to have faster inflation in the way you describe we would be hearing quite a lot from people on fixed incomes and we would probably also see the market drive up yields on government debt which could be something that would pose a burden for the taxpayer even if inflation did eventually come down.
Scott Sumner addressed this specific point in his post on Charles Goodhart:
[...] Goodhart is assuming debt markets care about inflation. They don’t. They care about NGDP growth. As long as NGDP growth is around 4%, long term nominal rates will remain relatively low. That’s the case regardless of whether the 4% NGDP growth is associated with o% RGDP growth and 4% inflation, 2% RGDP growth and 2% inflation, or 4% RGDP growth and 0% inflation.
The point about bond yields following NGDP not inflation seems reasonably clear in the data. Here’s the chart for the UK:
… in one simple graph, which I’ve aped from an excellent post by Nick Rowe and Stephen Gordon.
This shows the Bank of England’s performance in keeping CPI-CT growth close to a 2% trend growth path since the CPI target was adopted in 2004. The BoE has done even “better” on this metric than the Bank of Canada; there is not even the vaguest hint of a negative demand shock evident in this chart. The 2008 recession? It didn’t happen. 2011-12 recession? Nope. It’s not there.
And fiscal austerity? Are you kidding me? Even if you use the CPI-CT index as I have here, which ignores the VAT (and other indirect tax) changes, it’s not there. And even if you allow for “base drift” as the BoE did in 2008 due to the commodity price shock, the CPI level is still stuck well above a new 2% trend line starting in Summer 2008.
This is a graph of, if anything, an economy where demand growth is outstripping potential supply. Right?
Well, maybe not. Here’s “Why The Bank of England is Totally Wrong About Everything Being Just Fine”, in another simple chart, again aping Nick/Stephen’s graph:
So good… and then so awful. Who can possibly believe that the UK has deficient aggregate demand and also think that inflation targeting is a really good idea?