1 / 28

LECTURE 2: Section 1

LECTURE 2: Section 1. Micro/Macro Synthesis. The sad state of macroeconomic exchange rate research. Macro approaches have not yielded predictive models Meese & Rogoff ( JIE , 1983): nothing outperforms a random walk

abra
Download Presentation

LECTURE 2: Section 1

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. LECTURE 2: Section 1 Micro/Macro Synthesis

  2. The sad state of macroeconomic exchange rate research • Macro approaches have not yielded predictive models • Meese & Rogoff (JIE, 1983): nothing outperforms a random walk • “…there is remarkably little evidence that macroeconomic variables have consistent strong effects on floating exchange rates, except during extraordinary circumstances such as hyperinflations. Such negative findings have led the profession to a certain degree of pessimism visa-vis exchange rate research.” Frankel and Rose (Handbook of International Economics, 1995). • Pre-1970s goods market approach • Trade balance surplus (deficit) → currency appreciates (depreciates) • Post-1970s, Asset market approach • Financial assets, as well as goods and services matter • Markets are efficient and reflect all available public info • Prices jump immediately to new levels with no trading required • Homogeneous “representative” agents

  3. What’s different with the microstruture appoach? • Information not equally shared by all • Important private information exists • Price is “discovered” by market interactions • Heterogeneous market participants • Market makers, investors, central banks, corporates, … • May interpret and use common info differently • Hedgers and liquidity providers (speculators) • Different horizons • Trading mechanisms may affect pricing • Less transparency, the slower the price adjustment

  4. Importance of order flow • Signed volume determined by initiator of trade • There is always a match between buyer & seller in every trade, but price will be moved by initiating side of the market • Market makers will try to “read” your order when quoting & then react to any surprises • Macro models have no role for order flow • Macro fundamentals move prices with no trading required • Order flow conveys information regarding expectations of fundamentals • Confers advantage to large market makers with big clients • Traders update beliefs after observing order flow

  5. Why is there so much trading in FX? • Risk management: passing positions (Lyons’ “hot potato”) • Risk sharing is important to all • Pressure to manage positions results in certain taboos – for instance, the “drive by” • Don’t go to multiple counterparties simultaneously splitting a large order • Transaction taxes to reduce volume (Tobin tax) are ill advised • Uninformed market observers associate large volume with excessive speculation • Order flow conveys useful info, don’t impede this

  6. A representative macro, asset approach model • Asset approach to exchange rates: • The change in the exchange rate over a period like a year or month depends upon domestic & foreign interest rates, money supply, and other macro fundamentals • The residual reflects any microstructure effects like order flow • No empirical support

  7. A representative microstructure exchange rate model • Dealers determine prices as: • Where the change in the price between 2 transactions depends upon signed order flow, net inventory or positioning of dealers, & other fundamentals • Residual incorporates any public information, such as asset-approach effects

  8. Hybrid approach Incorporate both types of effects: • Aggregate micro variables to the time period associated with the macro model • Order flow is caused by information relevant to market participants • Order flow of hedgers less informative than that of speculators • Reflects updates of expectations

  9. LECTURE 2: Section 2 Macro Models of Exchange Rates

  10. Where we are heading The objective function for an active currency model: • It’s all about alpha!! • So how should we think about forecasting exchange rates? • Let’s start with some useful “building blocks”

  11. FX Foundations: Purchasing Power Parity • Think of the “law of one price” in terms of your international travels • Or rearranged for the exchange rate:

  12. FX Foundations: Uncovered interest rate parity The expected return from investing in a similar asset in two different countries should be equal: • The expectation is conditional on the info available at t By normalizing on the log of P (p) for approximate UIP, via forward iteration one can solve for the current exchange rate as the current expected future value and the sum of expected future interest differentials:

  13. Flexible-Price Monetary Model Assume PPP holds continuously & appeal to money market equilibrium & UIP (all variables in logs except for interest rates) • Today’s exchange rate depends on the discounted future value of fundamentals

  14. Sticky-Price Monetary Model • Also builds on money market equilibrium & UIP • But goods & services prices adjust slowly compared to financial asset prices (like exchange rates and interest rates) • So PPP does not hold in short-run & add a “wedge term” to the prior model to reflect deviations from PPP • As prices adjust to new equilibrium, the wedge term value falls • exchange rates more volatile than fundamentals and may “overshoot” • Exchange rates and interest rates must adjust “too much” to maintain money market equilibrium & UIP in short-run until the price level catches up

  15. Portfolio Balance Model Now there is imperfect substitutability between domestic & foreign assets • UIP is relaxed so FX risk premia may be earned (uncovered returns need not be equal on similar assets priced in different currencies) PPP is not assumed or necessary The exchange rate “balances” supply & demand for assets • Expected changes in exchange rates affect asset demands • The level of the exchange rate affects asset supply via the balance of payments (trade deficit (surplus)→asset outflows (inflows)) Asset demands depend upon the domestic interest rate & expected domestic currency return from foreign currency bonds • Changing asset supplies induce exchange rate changes to ensure all bonds held

  16. Portfolio Balance Model Model has 5 equations: • Money Demand • Domestic Bond Demand • Foreign Bond Demand • Foreign Bond Supply to domestic residents (trade balance + interest income) • Wealth constraint

  17. General Equilibrium Models Objective function is to maximize the representative individual’s utility • Utility f. may include money or there may be a cash-in-advance constraint to introduce role for money Relevant exchange rate is the “real exchange rate”: the nominal rate adjusted for price levels (=1 if PPP holds) • This relative price of domestic to foreign goods is found as being equal to the marginal rate of substitution in consumption of those goods GE models have not been successful at matching the empirical moments of exchange rates

  18. LECTURE 2: Section 3 Taylor & Taylor, “The Purchasing Power Parity Debate” MM: Is there a value signal lurking here?

  19. Why Should (or not) PPP Hold • Goods arbitrage is the “enforcer” for internationally traded goods • Enforce the “law of one price” • Services not arbitraged and subject to big price differences across countries • Barriers to arbitrage • Tariffs, transpost costs, nontariff barriers • Differentiated goods (use PPIs rather than CPIs to help mitigate) • Absolute PPP is strong form • Relative PPP is weaker (more likely to hold)

  20. PPP: a long-run proposition? • Authors show in figure 1 that over long-run absolute PPP held well for UK & US • Better for PPI than CPI • Can be significant deviations from PPP in short-run • Figure 2 shows another view of short-run vs. long-run for cross-section of countries & relative PPP • Annual data: no PPP • 29 year avgs: PPP • For PPP, the real exchange rate must mean-revert • A random walk rejects the hypothesis • Panel & long-run studies suggest HL of shocks of 3 years or more • So can one trade on such info?

  21. PPP: nonlinearities? • Transaction costs may result in appearance of random real XR movements until tcost threshold exceeded & arbitrage results in reversion to PPP level • Threshold effects, where reversion to PPP is increasing with the deviation from PPP could also arise from • central bank intervention • coordination of opinion on fundamental valuation of currencies

  22. PPP: shifts in fundamental value • Why should there be a constant PPP value? • Accounting for shifts may allow better view of speed of adjustment • Current account or net debtor/creditor status • Net debtor must run trade surplus to repay debt • Depreciating real exchange rate • Balassa-Samuelson effects • If law-of-one-price applies to traded goods and this is sector with technological progress, while non-traded goods are “traditional” industries • Rising wages in rich countries causes nontraded goods prices to rise in those countries so that the overall price index rises faster than in poor countries • Rich currencies appear overvalued and real exchange rate tends to appreciate

  23. LECTURE 2: Section 4 Alquist & Chen, “Conventional and Unconventional Approaches to Exchange Rate Modeling and Assessment” MM: Does anything work?

  24. Research Design • Random walk is benchmark to beat • Analyze 3 models of exchange rates • Quarterly data for US, CA, UK, Euro area, 1970-2004

  25. Sticky-price Monetary Model • Logs of money & real GDP & all are local-foreign differences • Estimated Dynamic OLS panel regression • Mixed results, some match priors, some evidence of long-run relationship

  26. UIP (& unbiased expectations) • UIP suggests that high interest rate currency should depreciate • Positive coefficient for interest rate differential • Estimation suggests short-horizon negative sign, long-horizon positive sign

  27. External Imbalances • Intertemporal approach to the current account: net debtors (creditors) should run surpluses (deficits) • Exchange rates part of adjustment via trade balance & net foreign asset revaluation; a debtor with a deficit should have a depreciating currency • Use Gourinchas-Rey (2005) model where • Xm=log export/log import; al=log assets/log liabilities; xa=log exports/log assets • nxa measures external imbalance & increases (decreases) with assets (liabilities) & exports (imports) • μis normalized weight • Estimated coefficients have expected negative signs, larger US external imbalance, dollar should depreciate

  28. Forecasts • Use rolling error-correction regressions out-of-sample • Evaluate forecasts using MSPE of random walk compared to forecast • UIP for recent sample & long horizon outperforms RW for CAD & EUR • Sticky-price monetary model not successful • External imbalances model has limited success at short horizon for CAD & GBP • All in all not very encouraging results • Consistent with most academic exchange rate forecasting papers • Active currency management not about forecasting time series of exchange rates • You win by getting the cross-section ordering correct

More Related