We provide a new structural interpretation of the relationship between the slope of the term structure of interest rates and macroeconomic fundamentals. We first adopt an agnostic identification approach that allows us to identify the shocks that explain most of the movements in the slope. We find that two shocks are sufficient to explain virtually all movements in the slope. Impulse response functions for the first shock, which explains the majority of the movements in the slope, lead us to interpret this main shock as a news shock about future productivity. We confirm this interpretation by formally identifying such a news shock as in Barsky and Sims (2009) and Sims (2009). We then assess to what extent a New Keynesian DSGE model is capable of generating the observed slope responses to a news shock. We find that augmenting DSGE models with a term structure provides valuable information to discipline the description of monetary policy and the model’s response to news shocks in general.

As this week’s NEP-DGE report was particularly long, I chose to highlight a second paper. This one feeds on the recent interest with the impact of news on business cycle. In this case, the focus is on the term structure of interest rates, an often neglected feature of economic fluctuations.

This entry was posted on Sunday, April 18th, 2010 at 5:23 pm and is filed under Uncategorized. You can follow any responses to this entry through the RSS 2.0 feed.
You can leave a response, or trackback from your own site.

9 Responses to News Shocks and the Slope of the Term Structure of Interest Rates

How did this paper make it to NEP-DGE? It is not general equilibrium, as the article this paper is based on, Smets and Wouters (AER 2007) assumes Calvo pricing (or rather, non-pricing).

M.H- The DSGE model in the second half of the paper is a general equilibrium model. Its not clear to me whether you simply object to a particular modeling choice (Sticky Prices), or perhaps you misunderstand the term ‘general equilibrium’. Perhaps you could clarify.

I think MH means that, outside of a knife-edge case, firms would never have chosen that exact timing of price changes (regardless of the underlying adjustment costs or rigidities) so you’re in effect making a variable that’s endogenous in theory be exogenous in your model.

Once you start doing that to endogenous variables you sort of lose the claim to “general equilibrium”, regardless of how popular this move is.

Agent Continuum: If you are going to make that argument, you need to apply it to all model features. For example, in Kydland-Prescott (1982, Time-To-Build and Aggregate Fluctuations) the investment technology is such that there are lags between when an investment project is started and when it becomes new capital. I think we would agree that “firms would never have chosen that exact timing of [investment lags]” (regardless of the underlying investment technologies). Despite this, I would still label Kydland-Prescott 1982 a general equilibrium model. Your logic can be applied to all parts of standard DSGE models. For example, technology is modeled as exogenous yet we know it is the result of agents choosing to produce new technology. Failing to model the technology development process explicitly does not mean we do not have a general equilibrium model.

Our models are by necessesity abstractions of reality. The important question is whether or not the modeling choices we make are reasonable abstractions of reality, for the question at hand. That last phrase is important when evaluating modeling choices. In our case, we believe the Calvo shortcut provides a reasonable description of aggregate inflation dynamics. Having a reasonable description of aggregate inflation dynamics is what we need to answer the key question in the paper: What is the impact of news shocks on the slope of the term structure of interest rates?

Sorry I helped derail the comments here in some other direction. I’m sure you’d rather see people commenting on the more substantive point of your paper!

sticky prices is a technology. Technologies do not apply to quantities only. A technology can be a quadratic cost of adjusting capital, or a quadratic cost of adjusting prices. Technology can be a 4 quarter lag on when investment become capital, or a 4 quarter contract on wages or 4 quarter delay before one can adjust prices. Technology can be randomly assigning people to work (indivisible labor) or randomly allowing firms to adjust prices (Calvo).

There are many legitimate criticisms of various specifications for nominal rigidities. However, claiming they are not general equilibrium is not one of them.

I have railed against Calvo pricing several times on my blog because in most models it is an unrealistic assumption. Indeed, results often depend critically on such price rigidities. Calvo pricing assumes that always the same proportion of firms changes prices, no matter what happens. This is patently wrong when there are news shocks. Firms will want to exploit news to adapt prices, thus the proportion of firms changing prices has to change. Whether this is significant, I do not know, but Calvo pricing does not give the flexibility to figure this out. And that is also why it is not general equilibrium.

Economic Logician: The conclusion you draw about Calvo prices not being general equilibrium does not follow the logic of your argument. Your argument is that the Calvo technology predicts unrealistic proportions of firm level price changes. That is, when solving for the equilibrium of the model the dynamics don’t match the data. This is a valid criticism of the model, but to call a model not general equilibrium because it has ‘unrealistic’ assumptions or doesn’t match the data along some dimension (once solved) defies logic.

In answer to the question of substance in your post. Firms in our model do change prices in the face of news. Our parameter estimates show that there is little price rigidity, and no indexation of prices to past inflation. That is, inflation is forward looking. This result appears consistent with your intuition.

How did this paper make it to NEP-DGE? It is not general equilibrium, as the article this paper is based on, Smets and Wouters (AER 2007) assumes Calvo pricing (or rather, non-pricing).

M.H- The DSGE model in the second half of the paper is a general equilibrium model. Its not clear to me whether you simply object to a particular modeling choice (Sticky Prices), or perhaps you misunderstand the term ‘general equilibrium’. Perhaps you could clarify.

-Chris

I think MH means that, outside of a knife-edge case, firms would never have chosen that exact timing of price changes (regardless of the underlying adjustment costs or rigidities) so you’re in effect making a variable that’s endogenous in theory be exogenous in your model.

Once you start doing that to endogenous variables you sort of lose the claim to “general equilibrium”, regardless of how popular this move is.

Agent Continuum: If you are going to make that argument, you need to apply it to all model features. For example, in Kydland-Prescott (1982, Time-To-Build and Aggregate Fluctuations) the investment technology is such that there are lags between when an investment project is started and when it becomes new capital. I think we would agree that “firms would never have chosen that exact timing of [investment lags]” (regardless of the underlying investment technologies). Despite this, I would still label Kydland-Prescott 1982 a general equilibrium model. Your logic can be applied to all parts of standard DSGE models. For example, technology is modeled as exogenous yet we know it is the result of agents choosing to produce new technology. Failing to model the technology development process explicitly does not mean we do not have a general equilibrium model.

Our models are by necessesity abstractions of reality. The important question is whether or not the modeling choices we make are reasonable abstractions of reality, for the question at hand. That last phrase is important when evaluating modeling choices. In our case, we believe the Calvo shortcut provides a reasonable description of aggregate inflation dynamics. Having a reasonable description of aggregate inflation dynamics is what we need to answer the key question in the paper: What is the impact of news shocks on the slope of the term structure of interest rates?

But time-to-build IS the technology. “Sticky prices” is not a technological constraint.

In G.E. it’s “fair game” to only take endowments, technology and preferences as exogenous (and maybe institutions), not prices or quantities.

Sorry I helped derail the comments here in some other direction. I’m sure you’d rather see people commenting on the more substantive point of your paper!

sticky prices is a technology. Technologies do not apply to quantities only. A technology can be a quadratic cost of adjusting capital, or a quadratic cost of adjusting prices. Technology can be a 4 quarter lag on when investment become capital, or a 4 quarter contract on wages or 4 quarter delay before one can adjust prices. Technology can be randomly assigning people to work (indivisible labor) or randomly allowing firms to adjust prices (Calvo).

There are many legitimate criticisms of various specifications for nominal rigidities. However, claiming they are not general equilibrium is not one of them.

I have railed against Calvo pricing several times on my blog because in most models it is an unrealistic assumption. Indeed, results often depend critically on such price rigidities. Calvo pricing assumes that always the same proportion of firms changes prices, no matter what happens. This is patently wrong when there are news shocks. Firms will want to exploit news to adapt prices, thus the proportion of firms changing prices has to change. Whether this is significant, I do not know, but Calvo pricing does not give the flexibility to figure this out. And that is also why it is not general equilibrium.

Economic Logician: The conclusion you draw about Calvo prices not being general equilibrium does not follow the logic of your argument. Your argument is that the Calvo technology predicts unrealistic proportions of firm level price changes. That is, when solving for the equilibrium of the model the dynamics don’t match the data. This is a valid criticism of the model, but to call a model not general equilibrium because it has ‘unrealistic’ assumptions or doesn’t match the data along some dimension (once solved) defies logic.

In answer to the question of substance in your post. Firms in our model do change prices in the face of news. Our parameter estimates show that there is little price rigidity, and no indexation of prices to past inflation. That is, inflation is forward looking. This result appears consistent with your intuition.