Exchange Rate Economics

1
Exchange Rate Economics

• John Williamson
• Senior Fellow, Peterson Institute for
  International Economics

Paper prepared for World Banks Growth
Commission, presumably concerned with whether
Asian model should be applied more generally
2
Exchange Rate Economics revolutionized since era
of flow models in 1960s
Now seen as forward-looking asset price
Current accounts influence comes only in pinning
down long-run value of RER (either uniquely by
PPP or by PPP plus real factors), with transition
modeled by assuming the representative agent is
rational and enjoys perfect foresight

• Exposition of conventional models
• Problems with these models and the most promising
  alternative
• Dutch disease
• Policy implications
• Questions for additional research

3
The Standard Model

• Long-run equilibrium characterized by current
  account imbalance such as to keep NIIP/GDP
  constant
• Level of equilibrium NIIP/GDP determined by
  factors considered in intertemporal theory of c/a
• Steady-state is independent of price level
• i.e. PPP prevails between equilibria
• either shocks are entirely monetary
• or real shocks are of second order of importance
• empirical evidence something close to PPP holds
  in L-R, but estimates of LRERs have significant
  coefficients also on NFAs, tot, G/Y, productivity.

4

• But should s-s be required as equilibrium
  condition?
• Doesnt this require that development be
  complete?
• Alternative allow NIIP/GDP, but nothing else, to
  vary
• Theory of debt cycle
• In the short-run
• UIP
• Overshooting
• Portfolio models
• Meese and Rogoff showed Dornbusch and Portfolio
  Models were outperformed by random walk in S-R
  (for industrial countries)
• Purists reject the L-R/S-R distinction and extend
  the S-R to L-R but impose a transversality
  condition with similar effect.

5
Problems

• Empirical implications of the standard model
• -- Exchange rates respond systematically to
  changes in the fundamentals (money supplies,
  income levels, interest rates, expected inflation
  rates, tot, productivityat least insofar as
  current values influence perceptions of permanent
  values), but do not change without news
• -- Money is made by having a good notion of the
  implications of the fundamentals, and will be
  lost by following chartist rules (Friedman)
• -- Exchange rate changes are normally
  distributed, no scope for bubble-and-crash
  dynamics

6

• All these implications are strongly counter to
  empirical evidence
• Meese/Rogoff study (still dominant conclusion,
  though Gourinchas and Rey argue that adding the
  exchange-rate induced change in wealth to the c/a
  permits a degree of S-R forecastability, and in
  Rio Carneiro and Wu cast similar doubt on the
  result for EMs) and what explains the DM/euro
  rollercoaster in 1993-2003?
• Chartist rules are used and profitable
• DM/ over 1986-95 had standard deviation of
  0.0029 per day but 3 days with changes greater
  than 0.015 (probability once every 7,000 years).
  Frequent apparent bubble-and-crash dynamics
• Standard model is a hopeless empirical failure
• But models are replaced only by a superior theory

7

• Candidate of de Grauwe and Grimaldi behavioral
  theory
• Same structure as that originally developed by
  Frankel and Froot in 1986, but more careful
  relationship to behavioral finance literature
• FX market populated by agents who use
  fundamentalist and chartist strategies
• Agent may act as either, and may change tactics
  in response to other rule proving more profitable
  (bounded rationality)

8

• No analytical solution, but several thousand
  simulations suggest
• That exchange rate changes are disconnected from
  fundamentals, though level is cointegrated with
  its fundamental value
• That chartist rule tends to be more profitable
  than fundamentalist one, though better is to
  switch
• That exchange rate changes have fat tails
• That the exchange rate is sometimes, but
  unpredictably, disconnected from its fundamental
  value and instead involved in bubble-and-crash
  dynamics
• That is, the model appears consistent with
  stylized facts.

9
Dutch Disease

• Is it a dangerous condition to be avoided by
  policy measures or welcomed as an improvement in
  a countrys situation?
• Dangerous condition
• Parable of export-led growth in E. Asia
• Possibility of interruption by Dutch disease
• Developed country has more chance of giving good
  living standard to all citizens (exceptions)
• Prudent act of investment (like Indonesia in 1978)

10

• Welcome condition
• Dominant view among economists
• Enlarges opportunities of domestic residents
• May involve painful S-R adjustment, e.g. in size
  of export sector
• Natural, efficient way to achieve this is by
  appreciation of RER.

11

• What does econometric evidence say?
• Evidence of a negative relationship between
  misalignment and growth (Razin and Collins)
• Evidence of such a relationship in developing but
  not developed economies (Prasad, Rajan, and
  Subramanian)
• Using a cross-country growth regression, Aguirre
  and Calderon get the usual results plus
  significant negative relationship between growth
  and misalignment driven largely by big
  misalignments (but the g-maximizing policy
  appears to be mild undervaluation)

12
An interesting result is that given change in RER
may have differential effect on growth if it is
an equilibrium phenomenon or a misalignment. Why?
Perhaps because of different judgments of the
private sector on the permanence of change. RER
decrease increases profitability of investing in
non-tradables to offset reduced profitability of
investing in tradables (sold on world market). No
reason to anticipate a crisis. Misalignment is
quite different in both respects.
13
If one concludes that Dutch disease is bad for
growth, what can policy do about
it? Conventional view little e not a policy
weapon because it floats even if it doesnt,
offset by induced inflation. But is this right?
Intervention is effective under the behavioral
theory of e (because it increases the rewards of
the fundamentalists). A potent instrument for
avoiding large misalignments by chartists jumping
on bandwagons. Aguirre-Calderon say no need to
worry about a real appreciation caused by a
permanent change.
14

• Difficult case is one of uncertainty as to
  whether a strengthening is temporary or
  permanent. Costly to make error.
• Is there another policy instrument?
• Variation in prudential regulations no, not
  cyclical variation.
• Capital controls
• Stabilization or Endowment Fund held outside
  country
• Taxes on entry of foreign capital as alternative
  to encaje, or on foreign interest income.

15
Policy Implications

• Assume that exchange rates are determined by
  behavioral finance model rather than standard
  model.
• Fixed e requires same 4 conditions
• Opt currency area
• Bulk of trade with partner country
• Macro policy consistency
• Institutional arrangements to ensure credibility.

16
Implication re floating should it be freeish or
managed? Standard model suggests free floating
rate normally close to fundamental value, and
there isnt much the authorities can do
anyway. Behavioral model challenges both
propositions. But does not endorse stable but
adjustable still crisis-prone. I.e. choice is
BBC, managed floating, unmanaged float. Band
main argument for adopting is to gain help of
speculators in stabilizing the rate. Requires
credible margins. Authorities have squandered
credibility, so no BBC. Unmanaged float
laissez-faire would be ideal if es behaved as
portrayed in standard model. But they dont.
17

• Two possible principles for more systematic
  management
• Leaning against the wind but why if the wind is
  blowing the right way? Possible answer because
  this will help fundamentalists make money.
• Reference rates. Need to secure agreement on ref
  rates. Political issue will countries agree to
  central rates despite traditional objection that
  all they can hope to identify are disequilibrium
  rates? (Perhaps limited obligation.) Also
  technical issue do the IMFs Macro
  Balance/External Sustainability approaches yield
  similar outcomes to ERER approach?

18

• Because of nature of obligation (not to push the
  rate away from ref rate), no conflict with IT
• Benefits
• Private market has sense of what official world
  believes equilibrium to be
• Basis for public debate about e
• International endorsement would aid in resisting
  cyclical appreciation caused by export boom or
  inflow surge. Most important for EMs/developing
  countries.

19
Additional Research

• Not really research, but importance of IMF
  publication of its figures on Macro Balance v.
  ERER Approaches.
• And understanding of whether Dutch disease is
  dangerous, and why.
• And?

20
Concluding Remarks

• Big step forward in 1970s when flow models were
  replaced by models that regard e as
  forward-looking asset price
• But next step has not followed it is overdue.
• Alternative model has important implications.
• Doesnt idealize laissez-faire.
• Or recommend fixing e, or reverting to adjustable
  peg, or abandoning floating or IT
• But it does require the IMF to accept the duty of
  negotiating reference rates and enforcing the
  obligations they would impose
• Which would at least help countries avoid
  misalignments that the official sector can see
  rest on temporary factors.