Backus @ NYU (research)

David Backus

Miscellaneous research papers

"Identifying Taylor rules," with Chernov and Zin, slides. Another outline of a paper we hope to write. The idea is to explore the possibility that forecasts, asset prices, and tight economic structure can identify the parameters of the Taylor rule and shocks to it. We think identification is less important, at this stage, than coming up with models that account for most of the observed features of bond returns and business cycles.
"International capital flows," with Cooley, Henriksen, and Moller, slides. We're working on a paper with a title something like this, but there's no ms yet. The current emphasis is on low-frequency movements in capital flows and their origins. Demographics is one possible driving force. Despite the lack of a paper, we've presented parts at a number of places.
"Sources of entropy in representative agent models," with Chernov and Zin, paper, slides. Forthcoming, JF. The paper is part methodology and part intuitive explanation for how popular representative agent models work. We propose two metrics for summarizing the properties of asset pricing models and apply them to representative agent models with recursive preferences, habits, and jumps. The metrics describe the pricing kernel's dispersion (the entropy of the title) and dynamics (horizon dependence, a measure of how entropy varies over different time horizons). Most popular models generate lots of entropy, but several have difficulty doing it without also producing too much horizon dependence. All of this depends on parameter values, so there's more to come on how to hit both targets.
"Monetary policy and the uncovered interest rate parity puzzle," with Gavazzoni, Telmer, and Zin, paper. There's a longstanding issue in international finance, that high interest rate currencies tend to rise in value. That gives investors a double kick (high rate, appreciation on average): what people have come to call the carry trade. Presumably the excess return from this strategy reflects risk of some sort. We build a relatively simple macro model in which the risk can be traced to monetary policy. It's an example more than a complete explanation, but suggests where you might look for a solution.
"Disasters implied by equity index options," with Chernov and Martin, paper, slides, Matlab programs. Published JF 2011. The possibility of large adverse events -- "disasters" -- can generate much larger risk premiums than you would get with normal (Gaussian) risks. The question is how much of this kind of thing is plausible. If you believe, as we do, that most risk reflects the economy as a whole, the difficulty in assessing the probabilities of extreme events is that they don't happen enough to allow precise estimates. We back the probabilities out of equity index options, which value such risks whether or not they occur in our sample. The resulting disaster probabilities are much smaller than those calculated by Barro and his coauthors.
"Cyclical component of US asset returns," with Routledge and Zin, paper, slides. We look at two old facts -- the stock market and the yield spread both lead the economy -- and note that they imply excess returns lead the cycle. We construct an exchange economy with similar features, a close relative of the Bansal-Yaron model in which changes in risk lead directly to changes in expected growth.
"Asset prices in business cycle analysis," with Routledge and Zin, paper, slides. First step toward a quasi-analytic solution of a business cycle model with recursive preferences. We know more now than we did then, will fix this up in some form in the near future.
"Taxes and the global allocation of capital," with Henriksen and Storesletten, paper, slides. Published JME, 2008. We see significant differences across countries in investment rates (ratio of investiment to GDP) and capital intensity (ratio of capital to GDP). We attribute some of this to differences in capital taxes. The main issue, here, is explaining why taxes based on revenue data are so different from those derived from the tax code.
"Cracking the conundrum," with Wright, paper. Published Brookings Papers, 2007. We interpret movements at the long end of the yield curve as reflecting, in large part, movements in risk premiums. It's an old story, but we bring some new evidence to bear on it.
"Current account fact and fiction," with Henriksen, Lambert, and Telmer, paper, slides, updated figures (US net worth, saving, investment). We never did anything with this, but it makes a useful point: that we're not really sure what international capital flows are telling us. Inflows could be signs of good news or they could be portents of disaster. As it turns out, disaster did occur, but in our view for much different reasons.
"Exotic preferences for macroeconomists," with Routledge and Zin, paper, slides, Matlab programs. Published in the 2004 NBER Macro Annual, but the link above has some corrections that were made after publication. The idea is to review recent advances in the theory of recursive preferences in user-friendly terms. Later papers with Zin and others developed loglinear approximations further.
"Discrete time models of bond pricing," with Foresi and Telmer, paper, NBER version, Matlab files - Self-Extracting. Published in a volume edited by former NYU colleague Bruce Tuckman that seems to have disappeared from sight. Bruce's fixed income textbook is available, though, and very good. Our paper is a review and synthesis of bond pricing models, including Vasicek, CIR, HJM, and many others. The idea is to describe these models in common language and simple terms, which we take to mean pricing kernels in discrete time.
"Oil prices and the terms of trade," with Crucini, NBER version. Published in JIE, 2000. A model of the impact of oil prices on business cycles, exchange rates, and capital flows. Mario is responsible for the novel parts, including the treatment of oil exporters.
"Accounting for biases in Black-Scholes," with Foresi, Li, and Wu, paper. We never got around to revising this one, but it has a nice result, mostly due to Liuren: that the slope and curvature of volatility smiles has a simple connection to the skewness and kurtosis, respectively, of the risk-neutral distribution. Based on a Gram-Charlier approximation.

Conference discussions