What should a central bank do when producer and consumer prices diverge?

In this post, Nick looked at this graph

and made the following comment:

In a counterfactual world, where the Bank of Canada was supposed to be targeting the level path of NGDP to follow that 5% trend line, what would we say? We would say that monetary policy was a little too loose in the years leading up to 2008, and then suddenly became much too tight in 2008, and stayed too tight thereafter.

In that counterfactual world, where the Bank was supposed to be targeting NGDP, but actually allowed NGDP first to rise a little above target, then suddenly fall well below target, we would all be looking at that third graph, and we would all be blaming the Bank of Canada for the 2008 recession.

(emphasis in the original)

There's something special about that episode, something I've mentioned before: producer prices (roughly approximated by the GDP deflator) grew much more quickly than consumer prices (roughly approximated by the CPI).

In a one-good closed-economy model where output is 'GDP', we don't make the distinction between producer and consumer prices. But in a multi-good economy with trade, it's often important to make this distinction. Starting in 2002, the prices of commodities rose relative to those of consumer goods, and you see this in the divergence between the GDP deflator and the CPI:

The GDP deflator increased by about 22% in log terms between 2002 and 2008, while the CPI increased by 16%.

This looks a lot like that staple of TV and film: the chase scene in which one persuer is chasing two people. The obvious strategy for those on the run is to split up: you can only follow one person. This forces the person chasing them to make the choice of which one to track down.

What should a central bank do when producer and consumer prices diverge? My (possibly incomplete) understanding is about how inflation affects welfare is on its effect on the price of consumption goods, especially in a world where nominal wages are slow to adjust. So it makes sense to me to make consumer prices the focus of attention and let producer prices go.

I'm not sure how to interpret this next graph, but I made it and I may as well post it. It compares NGDP with the series you get when you multiply real GDP by the CPI. I guess it's the counterfactual NGDP series for the scenario where producer prices and CPI stayed together:

The NGDP series suggest that monetary policy may have been too loose in the years leading up to the recession. But I think the better interpretation is that when producer and consumer prices diverged, the Bank of Canada - wisely, in my view - chose to track consumer prices and let producer prices go.

But it also raises the question of what happens to the volatility of consumer prices if we adopted NGDP targeting. If the Bank of Canada decides to chase down NGDP and let the CPI go where it may, this may introduce welfare costs that may or may not be offset by gains elsewhere. (I think this is what Felipe was getting at in his comment on Nick's post.)

I'm pretty sure the people at the Bank of Canada are running exactly these sorts of scenarios through TOTEM. I wonder what sort of results they're getting.

 

 

Comments

[Peter N]

There has also been a huge runup both in consumer debt (highest of all the developed countries) and housing prices (higher relative to long term trend than the recent US record).

Also, if unemployment tracks NGDP, there are welfare costs to tight control of inflation.

Then there's Canada's exposure to the US and Chinese economies.

[Nick Rowe]

I've been thinking through this, trying to get my head straight on what price level to target.

Take the extreme, and clearest, case. Suppose we export everything we produce, and import everything we consume.

Sure, we would like to stabilise our incomes, measured in terms of import prices in C$, because that's what we care about. But targeting the CPI might not do that, if our producer prices are sticky in C$, and if targeting the CPI would mean an exchange rate that made it impossible to sell our exports.

The usual prescription would be to target the stickiest price, and that isn't necessarily the CPI.

I need to keep thinking about this.

[RebelEconomist]

In my view, a central bank should ideally target a price index of everything traded against money. No hedonic pricing. Probably a Fisher index (although I have not really thought that one through). In short, Pbar in the equation of exchange. The idea being that the central bank should be supplying enough money to facilitate transactions but not so much or so little as to disturb the overall price level. A central bank can target consumer prices if it likes, and I think that would be sufficient to stabilise the overall price level in the long run, but, as we have seen, the danger is that asset prices can diverge so far that the central bank has to be really strong to keep its focus on consumer prices. In truth, I think that the choice of consumer prices was - perhaps unwittingly - a bit political: the punters like their central bank to hold down the price of what they buy, but not what they own.

[Determinant]

I'm always a radical when it comes to real estate prices: I want them to flatline or fall a bit because in the next ten years I want to buy a house.

I'm 30. I want a nice-enough house that I can look after, that doesn't leak that has enough room for the Missus (haven't found her, yet) and the two kids. My brother has moved into that stage which is why I'm an uncle. Rising real estate prices do not serve me one bit.

[Alex Godofsky]

"What should a central bank do when producer and consumer prices diverge?"

Target the path of NGDP, just like always :)

[Peter N]

It's odd that monetarists seem to be giving up quantity theory and the use of monetary aggregates at exactly the point where there are, for the first time, properly constructed aggregates. Doesn't this deserve some discussion?

NGDP is a coincident indicator, whereas the hope has been that aggregates would be leading indicators. I'd rather have leading indicators than faith in expectations channels and a rather elderly Chuck Norris.

Also maybe the answer isn't to target only one number. Perhaps, given the current array of tools, there could be 2 targets. The problems often seem to arise when one of the untargeted variables reaches extreme values.

[Peter N]

"What should a central bank do when producer and consumer prices diverge?"

Do you mean a central bank in general or the Bank of Canada? Having the much bigger US as a neighbor and major trading partner makes Canada a rather special case. Their NGDPs are highly correlated.

[Felipe]

I'm not sure this is the same as what I was commenting in the other post, but it certainly is related.

A central bank (of an economy with a large export sector) that chose to target total AD essentially surrenders domestic activity to the external demand of its exports. In other words, NGDP targeting seems to exacerbate the Dutch disease.

This can become a serious problem if your exports are volatile, but even in a low actual volatility environment it is problematic for commodity exporters because the supply is finite: it is anti-insurance.

[anon]

"My (possibly incomplete) understanding is about how inflation affects welfare is on its effect on the price of consumption goods, especially in a world where nominal wages are slow to adjust. So it makes sense to me to make consumer prices the focus of attention and let producer prices go."

Does inflation really affect welfare through consumer good prices, _once you adjust for supply-side conditions_? My answer is no: a nasty terms of trade shocks will raise inflation and make consumers worse off, but the central bank cannot do anything about it, and should not even try. So the answer is target producer prices, because they are a better approximation to monetary conditions (NGDP/total expenditure) that the central bank _can_ do something about.

[Frank Restly]

Nick,

"Sure, we would like to stabilise our incomes, measured in terms of import prices in C$, because that's what we care about."

Why is stabilizing income better than say stabilizing an asset to liability ratio? Income (or inflation adjusted income) is one way to measure wealth. Asset / Liability ratio is another. I fail to see why one measure should be "cared about" over another.

[Bubbles and Busts: Will NGDP Targeting Increase Consumer Price Volatility?]

This leads Gordon to beg the question, What should a central bank do when producer and consumer prices diverge?...

Gordon and Rowe’s focus is strictly on Canada, but given the hype surrounding NGDP targeting in the US, it might prove interesting to compare similar data.

[Nick Rowe]

Bubbles and Bursts has some good charts. Especially his chart comparing Canadian and US GDP deflators. US GDP deflator is much smoother, especially around 2008/9. That was also a time when the exchange rate moved a lot, IIRC.

[Andrew Baldwin]

As far as I know the only central bank that focuses on producer prices is the central bank for India, which monitors a wholesale price index. Here the reason seems to be mainly the existence of multiple CPIs for different segments of the population. If India had one CPI that covered virtually all households, their central bank would probably also target consumer prices.
I am not sure, even as a simplification, it is reasonable to treat the GDP deflator as a producer price index. After all, a large part of the deflator consists of personal expenditures, which are mostly deflated with consumer price indexes, and these consumer price deflators also occur in other places, such as calculating the value of physical change in inventories for retail trade. If one considers the GDP deflator as a producer price series, would one also consider the deflator for domestic final expenditures, for final domestic demand or for personal expenditures as producer price indexes? It seems to me that they should all be viewed as hybrid series, a mix of consumer and producer price index series, with consumer price series playing a subsidiary role in the GDP deflator but a dominant role in the personal expenditure deflator. As far as I know the only inflation targeting central bank in the world that targets a National Accounts deflator is the US Fed, and it targets the personal expenditure price index.
The more interesting question is whether a central bank should monitor consumer prices on a national or a domestic basis. Assume as Peter N does that all consumer expenditures by the resident population are imports and all production is exported. Targeting a national-basis consumer price series, appropriate to escalation of the population’s incomes, would be a weighted average of consumer price series for the foreign countries in which residents visited or worked. It would be an absolutely hopeless series for the central bank to target because it would have so little influence on it. Domestic restaurant prices, domestic rents, domestic hotel rates, would have no place in it since by assumption all spending on these would come from non-residents. On the other hand, targeting a domestic-basis consumer price series, while quite inappropriate to the escalation of the population’s incomes, would be appropriate for the central bank, since it could control domestic prices.
The UK CPI monitored by the Bank of England and the Monetary Union Index of Consumer Prices both adopt the domestic approach in their calculation. So, for example, foreign students tuition fees in the UK are part of the UK CPI, but not UK students’ tuition fees paid abroad. The Canadian CPI, like most CPIs adopts a national approach in terms of its weighting but a domestic approach in terms of its pricing. The expenditure weights for tuition fees represent all tuition fees paid by students who are Canadian residents at schools within Canada and abroad, but only tuition fees at Canadian universities are priced.
It is hard for someone to get their head around including foreign students’ tuition fees in a target inflation indicator, which seems so contrary to any welfare interpretation of a consumer price index. However, suppose the whole world consisted of Mexico, the US and Canada, each with their own inflation targeting central bank. If they all targeted domestic-based consumer price series, everyone’s consumption would be targeted by one central bank or another; no consumption would be ignored. Not so much potential output would have to be sacrificed to meet the world inflation target precisely because each central bank was targeting the prices that they had best control of, but the inflation outcomes would be the same.
Governor Carney in his speech in New York City on February 24 actually gave as a key argument against NGDP targeting that the deflator was based on the domestic approach, not the national approach! Now that he is about to become governor of a central bank with a domestic-based inflation indicator, it is about time he got up to speed on this issue.

[Nick Rowe]

Andrew Baldwin has a good comment stuck in the spam filter.

[Determinant]

Mostly because I don't use Twitter:

I hope Stephen is reading this. I will be a delegate to the NDP Convention for my riding.

The Observer cost is a political donation, you get a tax credit of $625.

Therefore your effective observer fee is $575 for the three days. I'd love to meet if you want to come.