Good framework for micro to macro: distribution of units across this space

|
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
 7 views
of 27

Please download to get full document.

View again

Description
Comments on: “Investment and Employment Dynamics in the Short-Run and the Long-Run”, Avinash Dixit, OEP (1997) Why I chose this paper: Very clearly written paper which pushed the intuition and relegates the details to other texts
Share
Transcript
Comments on: “Investment and Employment Dynamics in the Short-Run and the Long-Run”, Avinash Dixit, OEP (1997)
  • Why I chose this paper:
  • Very clearly written paper which pushed the intuition and relegates the details to other texts
  • These models of adjustment costs in multiple dimensions are important in aggregate setting as the generate distributions and interesting responses (as we’ll see). EG:
  • Prices
  • Employment
  • Consumption (housing)
  • Share trading
  • Comments on: “Investment and Employment Dynamics in the Short-Run and the Long-Run”, Avinash Dixit, OEP (1997)
  • Good framework for micro to macro:
  • distribution of units across this space
  • shifts in this distribution (1st and 2nd moment aggregate shocks)
  • shifts in the thresholds, regulations, expectations etc
  • Nick BloomInvestment and UncertaintyWhat are the real options models predictions?
  • Investment and hiring/firing will be lumpy:
  • Evidence from Micro Data (Davis and Haltiwanger, 1992)
  • Investment and hiring/firing will respond slowly to shocks:
  • Wide range of evidence from micro and macro data
  • Investment and hiring/firing response effected by adjustment costs
  • No direct evidence that I know of (Bertola & Bentolila, 1990)
  • Investment and hiring/firing response effected by uncertainty:
  • Will discuss this today…
  • John Leahy and Toni Whited (1996)“The Effects of Uncertainty on Investment: Some stylized facts”Journal of Money, Credit and BankingOverview
  • Panel estimation of the effect of uncertainty on investment
  • The fundamental idea was to:
  • Use yearly volatility of daily returns of stock as a measure of uncertainty (σi,y≈ Σd(ri,y,d2) where ri,y,d is the stock returns for firm i, in year y on day d)
  • Estimate yearly firm investment as a function of this
  • Use firm and year controls to try and deal with other ommitted variables
  • Use GMM to try to deal with exogeneity
  • An important paper:
  • First general test of investment-uncertainty literature(ii) Good econometrics – GMM and panel estimation
  • Used interesting data – firm returns variance
  • Many cites - the rewards for being first…IBasic results
  • Uncertainty reduces investment
  • Covariance has no impact
  • Tobin’s Q controls removes uncertainty effects, but note:
  • Tobin’s Q also not significant
  • Tobin Q = Marginal Q both PC & CRS hold, real options assumes either PC or CRS does not hold
  • Message is no large direct impact of uncertainty (at yearly frequency)
  • Good paper – how could you build on this:
  • Theory – take a structural approach. No link with theory
  • Use Dixit & Pindcyk (1994), Abel & Eberly (1996)
  • Identification – everything moves together so not clear what is identifying what (GMM is no magic bullet)
  • Data – better measures of uncertainty, for example use implied volatility (forward looking) or a direct measure
  • Luigi Guiso and Giuseppe Parigi (1999)“Investment and Demand Uncertainty”Quarterly Journal of EconomicsOverview
  • Paper also estimates the effect of uncertainty on investment
  • Contribution is a great uncertainty measure – a survey of firms distribution of demand growth expectations
  • Used to generate a mean and variance of expected demand
  • So can estimate effects of variance controlling for the mean
  • The authors basically persuade the BOI to run this – ingenious!
  • Also had a range of tests of other implications of the real options theory – so much closer to the theory
  • The rewards for ingenious data organisation…Basic setupThe paper is heavily (and appropriately) orientated around the description of the empirical uncertainty measure One important innovation is they changed the specification for the uncertainty from levels to an interactionAlthough they do always include levels of uncertainty as a control [Note: whenever estimating interactions you must always include the levels of all variables as well, otherwise misspecified!]They find a “cautionary effect” of uncertaintyThis is obviously too small….I put this up to show you the QJE table format – full notes at bottom of the table. Always do this in your papers (including in drafts)They find a “cautionary effect” of uncertaintyOther parts of the paper
  • The paper also has other tests of the real-options, finding for σ:
  • More impact for irreversible capital (directly measured and indirectly measured using Shleifer and Vishny (1992))
  • More impact at higher market power
  • Occurs in structures, machinery & vehicle investment
  • Look reasonable and good to push the empirics with these additional tests – although small sample size (500 obs) so tenuous
  • My guess is these were – appropriately – requested during the refereeing process
  • Message is uncertainty reduces responsiveness
  • Good paper – how could you build on this:
  • Theory – take a more structural approach. Again no real link between theory and empirics
  • Identification – cross-sectional data without any IV
  • “So what” factor – push some interesting implication if uncertainty varies by time, region, industry etc..
  • Nick Bloom, Steve Bond & John Van Reenen (2007)“Uncertainty and Investment Dynamics”Review of Economic StudiesOverview
  • Paper also estimates the effect of uncertainty on investment…
  • Contribution is goes directly from theory to empirics to derive robust predictions and test these
  • Shows higher uncertainty generates a large and robust “caution effect”: 1 SD increase in σ ½’s investment response
  • Shows this is robust to aggregation (cross-section & time)
  • Takes a matched theory-empirics approach (panel, GMM)
  • Also has a range of tests of impacts of different types of adjustment costs, uncertainty effects and functional forms
  • The rewards of impressive modellingBasic setup
  • The paper is split between theory and empirics
  • Theory is simulation based (this is possible, but put code on-line):
  • Solves for time varying σ
  • Shows σ induces “caution” – like an adjustment cost
  • Shows this is robust to aggregation – empirically important
  • Shows on simulated data this effect can be estimated (by GMM)
  • Shows on actual data find similar results (by GMM)
  • Show robust to various tests (correctly requested by referees)
  • The “cautionary effect” of uncertaintyEstimated by GMM on simulated dataEstimated by GMM on actual dataUse empirical results to show variations in σ matterMessage is theory and empirics show uncertainty reduces responsiveness
  • Good paper(?) – how could you build on this:
  • Identification – everything moves together so not clear what is identifying what (GMM is no magic bullet).
  • Find some exogeneous shift in σ which does not effects levels?
  • Data – better measures of uncertainty, for example use implied volatility (forward looking) or a direct measure. Also could use higher frequency data (quarterly Compustat data)
  • Other outcomes – look at labor, R&D, productivity etc…
  • I think this is an interesting research area as uncertainty also looks like it is counter-cyclicalShaded areas are recessions
    Related Search
    We Need Your Support
    Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

    Thanks to everyone for your continued support.

    No, Thanks