Many central bankers struggle these days to understand what is the impact of globalization on inflation. Most people would agree that it helped to lower inflation over the last decade, but there is no consensus on what was the magnitude and whether it will last. There is a very good summary of these issues in the recent Jeffrey Frankel paper. See key points and my comments below.

  1. China lowers US inflation permanently, because steadily rising trade integration (as measured by trade/GDP) would lead to steadily rising real income, which would lead to lower inflation. Why? If you want to achieve a desired level of consumption then Wal-Mart bringing to you cheap Chinese goods reduces your demand for nominal wage increases. Problem with this argument is that if one excludes top 1% in the US income distribution the real incomes of US houselholds have fallen.
  1. Increased exposure to international competition could leads to increased competitiveness of domestic economy, productivity advances faster, so unit labor costs are lower and inflation is lower. This hypothesis is consistent with change in productivity trend in the United States that was first notice by Alan Greenspan in mid-1995 and once enough years of data where accumulated was formally model-tested (see Blinder, Reis 2005)

  2. There is an argument by Grossman and Helpman (Grossman, Gene, and Elhanan Helpman, 1991a, Innovation and Growth in the Global Economy, MIT Press: Cambridge. Grossman, Gene, and Elhanan Helpman, 1991b, “Trade, Knowledge Spillovers, and Growth,” European Economic Review 35, no. 2-3, April, 517-526. ) that even a one-off increase in country openness could lead to a permanent or long-lasting increase in productivity. The channel that operates here is that in XXI century productivity gains are increasingly based on innovation. Before companies grew by building new factories and employing new people. In the XXI century companies grow by inventing new products, and these products get produced in the world (one part here in house, one part there offshored, and usually assembled in China). If you want an extreme case of ideas-driven-growth think about YouTube, invented one and a half year ago in a garage by two computer geeks and sold yesterday to Google for US$1.5bn. So interacting with other countries in a form of trade in goods, services and ideas leads to higher pace of creation and absorption of innovation. Again shortly after Google Answers was proposed by Google, its modified Polish version was created in no time, offering various new functionalities to Polish Internet users. There are numerous papers providing evidence that supports this view (those with subscription can visit McKinsey Global Institute for a nice selction of papers on the issue of ideas-driven-growth).

  3. If these arguments were to be true then one would have expected that falling pricing-power of firms would lead to lower margins and to lower profits, while we observe the opposite in the data. The share of corporate profits in GDP rose to all-time high, so one needs to explain why companies facing tougher competition are able to deliver higher profits. Jacob Frankel posits, that maybe globalization led to higher increase in competition in the labor market than in the product market. Many financial sector economists would probably agree (some call it global labor arbitrage) or think about 1 million (5 percent of active labor force) Poles that left Poland and took better paid jobs in other EU countries. However, I would raise some issues here. Labor mobility is lower than goods mobility, for a Chinese it is probably easier to sell widgets in US than obtain working permit there, or for example she may choose to sell widgets on EBay, while the technology to travel via Internet is yet to be developed (certainly recent Rand report on innovations in 2020 does not mention such possibility). So it must be something else that works here. I think that Grossman and Rossi-Hansberg paper (2006, Jackson Hole) comes to the rescue, that trade today it is no longer wine for cloth, in XXI century we trade tasks, i.e. we trade parts of business processes. While by any measure the scope of offshoring and outsourcing is still small (see evidence collected in Rybinski book on globalization – in Polish) , it may have affected expectations of workers, who now see a pressing need to become more productive, more flexible in order to keep her job.

New view – flatter Phillips curve

Frankel also nicely summarizes the ongoing debate on the shape of the Phillips curve. But let me first recall recent Bank of England Governor Merving King speech on this topic in June this year. He presented UK inflation/unemployment data and concluded that relationship between unemployment and inflation changed from vertical in 1970-80ties to horizontal. In 1990-nowadays. Vertical Phillips curve means that generating extra inflation did not produce higher growth and did not reduce unemployment, real economy was unchanged and the only lasting result of allowing for higher inflation was …. higher inflation. In UK in recent 20 years we see remarkably stable and low inflation and unemployment falling from 10 percent to 5 percent. So it appears that the Phillips curve tripped and fell sometime around 1990s. King listed three factors behind the horizontal Phillips curve:

  1. structural reforms of the labor market in the UK
  2. improvement of terms-of-trade, falling process of goods (UK imports) relative to prices of services (UK exports). Higher purchasing power of take-home wage has allowed business to hire more people without the need to raise wages
  3. Migration to UK which raised the UK potential output

Now let me long-quote what Jacob Frankel wrote above the Phillips curve debate:

” …Much of the most recent thinking focuses on the slope of the Phillips Curve: the magnitude of the increase in inflation resulting from a given expansion of domestic demand (or the fall in inflation resulting from a given contraction in demand). Some suggest that globalization impliesthat inflation is less sensitive to domestic demand conditions, and more to global demand conditions, than it used to be: Borio and Filardo (2006), Fisher (2005), IMF (2006, pp. 106-108), Kohn (2006), and Yellen (2006), who calls this the “new view.” The argument is that foreign supply is more readily substituted for domestic output than before, so that the Phillips curve is flatter. (Firms have “less pricing power.”) Others suggest that globalization has produced a steeper Phillips curve: Dornbusch and Krugman (1976, pp. 570-573), Romer (1993), Rogoff (2004). The argument is that it is harder to raise output — a country pays the price of monetary expansion more quickly, especially if the exchange rate is floating — because the economy more closely approximates the frictionless perfectly competitive neoclassical paradigm. As much as international competition, Rogoff (2004) has in mind domestic sources of increased competitiveness from deregulation, privatization, decreased union power, and the advent of Wal-Mart, Amazon and EBay. Remarkably, both camps, those who argue that globalization makes the Phillips curve flatter and those who argue that it makes it steeper, are suggesting that it results in lower inflation. In the “new view”, a given monetary expansion, or a given target in terms of output, is associated with lower inflation. The Romer (1993)-Rogoff (2004) claim that the Phillips curve is steeper of course recognizes the implication that a given monetary expansion will lead to higher inflation. But it goes on to point out that precisely because it would accomplish little, central banks in highly open economies will refrain from monetary expansion. People are aware of this, which reduces their expectations of inflation. The result in the general equilibrium of rational expectations (Barro-Gordon, 1983) is that open economies will exhibit less inflationary bias than less open, less competitive economies. The attractiveness of this model from a theoretical viewpoint is that it provides a rationale for an increase in the level of globalization producing a permanent fall in the average rate of inflation. Romer (1993) and Lane (1997) produced evidence that more open countries indeed have lower inflation rates…”.

I have very little to add to this. However I think that economics profession should now deal with the following issue. Inflation is becoming a global phenomenon (see for example Ciccarelli, Mojon ECB paper ). If this is the case that all central banks collectively can control inflation and their collective easing/tightening results in tighter/easier global monetary condition, and if anything global liquidity was found an important variable explaining global inflation (be careful, do not implicate that the same holds for country level inflations, because it does not). However whenever there is a collective action that is associated with a cost (and stabilizing inflation is costly) then there is a temptation to shift costs to other actors, or an incentive to free ride. This issue has so far not been addressed by the economic profession.

I am finishing this post on the plane on my way back from National Bank of Belgium seminar , unfortunately I had to skip the second day of the seminar because I was invited to speak at the Polish Lisbon Agenda Forum tomorrow. Being a central banker involved in lots of administrative issue I find it hard sometimes to keep up with recent advancement in the economics profession. So I enjoyed immensely several excellent presentations, not to mention that one of the discussants, Jerzy Konieczny is an old friend of mine that I have not seen for a long time. I think that the paper by Olivier Blanchard and Jordi Gali is a very important contribution, but I particularly benefited from Marvin Goodfriend discussion, who showed what are the key contributions.

Let me summarize the key policy messages that – as I understood – can be derived form B-G paper. If I got something wrong hopefully somenone will correct me commenting this post.

  • New Keynesian model (or New Neoclassical Synthesis, for backgroud see Fed Richmond NNS Primer ) has become the dominant and standard tool to analyze monetary policy and derive policy recommendations. The policy recommendation coming from the standard New Keynesian model was that a central bank should focus on stabilizing inflation and that it resulted in an optimal outcome
  • The key problem with the New Keynesian model was that there was no unemployment in the model, while in practice central banks worry about both inflation and unemployment, thisnk about the Federal Reserve and its dual mandate to keep inflation low and to keep economy at full employment. There are indeed very few central banks in the world, if any, that pursue strict inflation targeting, although early years of the National Bank of Poland experience with the direct inflation strategy can probably be seen that way.
  • B-G model adds unemployment dimension to New Keynesian model by marrying two strands of the literature: New Keynesian model with Diamond-Mortensen-Pissarides model which is used for the analysis of the labor market frictions, labor market dynamics, and implications of alternative policy interventions on unemployment
  • Describe key features (will be added later)
  • This modification has very powerful consequences for the monetary policy conduct. Now focusing solely on stabilizing inflation results in very high unemployment volatility, which increases the central bank loss function (as central bank does not like inflation volatility and unemployment volatility, or more precisely they do not like inflation to move away from the target and unemployment to move away from full employment.
  • As Goodfriend showed, rough calibration of the model shows that central bank focusing on price stability when faces with negative 1 percent productivity shock would stabilize inflation and would produce much sizeable rise in unemployment (by few percentage points). However if a central bank conducted an optimal policy (minimizing loss) it would see unemployment rate rising only by 1 percent at the cost of inflation higher by 0.6 percent. Well, a pretty good deal, welfare improving indeed.
  • The model is small and tractable and one can safely bet that many extentions will mushroom, in particular many assumptions that are external (exogenous) to the model will e made a function of model structural parameters. I doubt whether it would change policy prescriptions.
  • This discussion comes at a very good moment indeed, as many ponder whether Federal Reserve with Ben Bernanke at the helm will formally adopt inflation targeting as its monetary policy strategy. The paper seems to justify the dual mandate and dual target approach, but I do look forward to central bankers discussion on this issue. I worry though, that this paper policy implications will be misunderstood by politicians who will call on central banks to target low unemployment.
  • My understanding of the model policy implications is this. Flexible inflation targeters should continue to do their job, making sure that the loss functions they choose is a close approximation of the social loss function. I also think that we are witnessing a huge positive productivity shock called globalization, which will (hopefully) last and probably accelerate in the coming years (unless protectionism surfaces again, as it did in post-war year). When B-G model is hit by positive productivity shock it tends to show deflationary outcomes amid falling unit labor costs. Of course the model does not have high oil prices driven by large China demand, but its clear and intuitive structure could also contribute to the debate on good deflation.

I think that the short answer to the question posed in the title of this post is this: central bankers should do their job keeping inflation low and stable. But after listening to discussion on this issue over the past few months I have a feeling, that this task will become symmetrically more difficult.