Exchange-Rate Determination, Parity Conditions and Forecasting

Introduction and context

Exchange rates are prices that connect national economies. They influence the domestic-currency cost of imports, the foreign-currency revenue of exporters, the value of overseas assets and liabilities, inflation transmission, international competitiveness and the required return on cross-border investment. For an international business, an exchange-rate movement is not merely a treasury concern: it can alter sourcing decisions, transfer prices, investment appraisals, debt-service costs and the viability of an entire market-entry strategy.

This chapter examines why exchange rates move and how managers can forecast them. No single model explains all exchange-rate behaviour across all horizons. Parity conditions provide economically disciplined benchmarks; monetary and portfolio-balance models explain macroeconomic forces; and forecasting practice combines fundamentals, market prices, technical evidence, econometrics and scenarios.

Learning objectives

By the end of this chapter, the reader should be able to:

1. Define exchange-rate quotations and interpret appreciation and depreciation.
2. Apply purchasing power parity, interest-rate parity and the Fisher effect.
3. Distinguish covered and uncovered interest parity and understand their assumptions.
4. Explain expectations, monetary and portfolio-balance models of exchange rates.
5. Construct and evaluate fundamental, technical, econometric and scenario-based forecasts.
6. Diagnose model limitations and translate forecasts into managerial decisions.

1. Exchange-rate conventions and market foundations

An exchange rate must be defined carefully. Let S_t denote the spot exchange rate quoted as units of domestic currency per unit of foreign currency. For example, S_t = ₹90/US$ means that one US dollar costs ₹90. An increase in S_t is a depreciation of the domestic currency; a decrease is an appreciation.

The reciprocal quotation is foreign currency per unit of domestic currency:

The reciprocal is valid only when the two quotations use exactly reciprocal units. Forward rate F_t,T is the price agreed today for delivery of foreign currency at time T. The forward premium or discount on the foreign currency, annualised over a period of length τ years, is commonly measured as:

The approximation using logarithms is:

The annualisation convention must be stated; market conventions differ by currency and instrument. Exchange rates are determined in decentralised foreign-exchange markets by orders from commercial banks, corporations, asset managers, governments, hedge funds and other participants. Spot prices respond to current transactions, while forward prices incorporate interest rates, risk premia, liquidity and expectations.

A currency’s nominal exchange rate should be distinguished from its real exchange rate. With domestic price level P and foreign price level P^*, a real exchange-rate index consistent with the quotation above is:

A rise in q_t means that foreign goods become relatively more expensive in domestic-currency terms under this convention. If the price indices are measured in comparable units, q_t is dimensionless. Interpretation depends on the quotation and index definitions, so analysts must avoid importing conclusions from an alternative convention without adjustment.

2. Purchasing power parity

Purchasing power parity (PPP) links exchange rates to relative price levels. Absolute PPP proposes that identical tradable goods should have the same price after conversion into a common currency, assuming frictionless trade, identical goods and no transport costs, tariffs, taxes or market segmentation. In aggregate form:

Here P_t is the domestic price of the representative basket and P_t^* is the foreign price. This strict version is rarely observed because consumer-price baskets contain non-tradable services and differ across countries.

Relative PPP is more useful in practice. It predicts that the rate of change in the exchange rate approximately equals the inflation differential:

where π and π^* are domestic and foreign inflation rates over the same interval. In logarithmic form, with s = ln S:

The approximation is appropriate for moderate changes. The exact discrete relationship implied by a price-level version is:

for constant inflation over the period. Thus, if domestic inflation is 8% and foreign inflation is 3%, the predicted domestic-currency depreciation is approximately 5%; the exact factor is 1.08/1.03 = 1.0485, or 4.85%.

PPP is a long-run benchmark rather than a reliable short-run trading rule. Deviations arise from transportation costs, tariffs, quotas, non-tradable goods, quality differences, pricing-to-market, capital flows, productivity differences and slow price adjustment. The Balassa–Samuelson argument further suggests that rapid productivity growth in a country’s tradable sector can raise wages and non-tradable prices, producing a real appreciation even without excessive nominal appreciation.

3. Interest-rate parity

Interest-rate parity connects spot and forward exchange rates with interest rates. It is based on the principle that comparable investments should have equal returns after exchange-rate conversion, subject to the relevant hedging and risk assumptions.

Covered interest parity

Covered interest parity (CIP) applies when the future exchange rate is hedged with a forward contract. With annual effective domestic and foreign interest rates i and i^*, the exact no-arbitrage relation for the domestic-currency-per-foreign-currency quotation is:

For continuously compounded rates r_d and r_f:

The variables must refer to the same maturity, compounding convention and credit quality. The relation assumes that borrowing and lending, currency conversion and forward trading can occur at sufficiently similar prices, and that capital controls and transaction costs do not prevent arbitrage. In liquid major-currency markets CIP is usually a powerful benchmark, although collateral, balance-sheet costs, funding liquidity and regulatory constraints can create persistent deviations.

Worked example. Suppose the spot rate is ₹90/US$, the one-year Indian interest rate is 7%, and the US rate is 4%. The theoretical one-year forward rate is:

The dollar trades at a forward premium of approximately 2.884% against the rupee, while the rupee trades at a corresponding forward discount under the reciprocal quotation. This is a pricing relationship, not a prediction that the future spot rate must equal ₹92.596.

Uncovered interest parity

Uncovered interest parity (UIP) compares expected, rather than contracted, future spot rates. In approximate form:

Equivalently:

UIP assumes risk neutrality, rational expectations and negligible risk premia, or alternatively that any risk premium is zero. In reality, investors may require compensation for currency risk, liquidity risk and sovereign or convertibility risk. Empirical evidence also documents the forward-premium puzzle: high-interest-rate currencies have not consistently depreciated by the amount predicted by UIP and may sometimes deliver positive excess returns. Consequently, CIP is primarily a hedged pricing condition, whereas UIP is a behavioural and expectations hypothesis.

4. The Fisher effect and expectations theories

The Fisher relationship separates nominal interest rates into expected real returns and expected inflation. The exact relation is:

where i is the nominal interest rate, r the expected real interest rate and π^e expected inflation. Therefore:

For moderate rates, the approximation i ≈ r + π^e is convenient but not exact. The Fisher effect states that, if expected real rates are stable across countries, differences in nominal interest rates reflect differences in expected inflation:

Combining the Fisher effect with UIP produces a link between expected depreciation and expected inflation differentials, broadly consistent with relative PPP. Each link relies on assumptions that may fail, including stable real rates, rational expectations, open capital markets and comparable risk.

Expectations theories explain how beliefs enter prices. Under rational expectations, agents use available information without systematic forecast errors, although they can still be wrong because of genuinely unexpected shocks. Adaptive expectations adjust gradually from past errors. Market-based expectations can be extracted from forward rates, but a forward rate generally equals the expected future spot rate plus a risk premium:

This additive expression is a simplifying representation in exchange-rate units; multiplicative or log forms may be preferable for proportional analysis. The risk premium is not directly observable and may vary with global risk appetite, fiscal credibility and market positioning.

5. Monetary models of exchange-rate determination

The monetary approach treats the exchange rate as the relative price of two monies. A simple money-demand function is:

where lowercase letters denote logarithms of money supply m, price level p, and real income y; i is the nominal interest rate; and α, β > 0. Money demand rises with income and falls with the opportunity cost represented by the interest rate.

Under purchasing power parity and equal money-demand parameters in two countries, subtracting the foreign equation from the domestic equation gives a flexible long-run exchange-rate relationship:

The signs reflect the quotation S as domestic currency per foreign currency. Holding other variables constant, greater domestic money supply tends to depreciate the domestic currency; greater domestic real income tends to appreciate it through increased money demand; and a higher domestic nominal interest rate tends to depreciate it in the static monetary model because it reduces money demand. This last implication can conflict with short-run interest-rate interpretations because interest rates may rise in response to inflation expectations or monetary tightening.

The overshooting model resolves part of this tension. Financial markets and exchange rates can adjust rapidly, while goods prices are sticky. An unexpected monetary expansion may initially cause the currency to depreciate more than its eventual long-run PPP level. As prices gradually rise and real balances fall, the exchange rate can appreciate from its initial overshoot toward the long-run value. The model highlights the importance of adjustment speed and expectations rather than asserting that every monetary expansion produces the same path.

Monetary models are most useful for long-run analysis and economies where money, inflation and policy credibility are central. They perform poorly when money demand is unstable, financial innovation changes monetary aggregates, capital flows dominate goods-market forces, or PPP deviations are persistent.

6. Portfolio-balance and asset-market models

The portfolio-balance approach recognises that domestic and foreign assets are imperfect substitutes. Investors allocate wealth among domestic bonds, foreign bonds and equity according to expected return, risk, liquidity and wealth. A stylised demand function for foreign assets can be written as:

where W is wealth, E[R_f and E[R_d] are expected returns, σ_f and σ_d are risk measures, ρ is return correlation, and L_f and L_d represent liquidity. The function increases with foreign expected return and wealth, but its exact response to risk depends on portfolio preferences and covariance.

A government deficit financed by issuing domestic bonds can increase the supply of those bonds held by the public. If domestic and foreign assets are not perfect substitutes, investors may require a lower domestic-bond price, a higher yield and a currency adjustment to hold the expanded supply. Conversely, official reserve purchases, quantitative easing, pension rebalancing and safe-haven flows can affect exchange rates even when current money growth or trade flows appear unchanged.

Portfolio-balance models are particularly valuable for analysing risk premia, sovereign debt, global liquidity and sudden changes in investor sentiment. Their weakness is empirical complexity: expected returns, portfolio preferences and risk exposures are difficult to measure, and causality can run in both directions.

7. Forecasting exchange rates

Forecasting should begin with a clearly specified horizon, quotation, purpose and loss function. A treasury manager hedging a three-month payable has a different requirement from a strategy team evaluating a ten-year foreign investment. The relevant question is not simply “What will the exchange rate be?” but “Which decision improves when uncertainty is represented explicitly?”

Fundamental forecasting

Fundamental analysis examines variables suggested by economic theory:

• inflation and inflation expectations;
• interest-rate differentials and central-bank policy;
• money and credit growth;
• fiscal deficits, public debt and external balances;
• terms of trade and commodity prices;
• productivity and competitiveness;
• capital flows, reserves and external financing;
• political, institutional and sovereign-risk conditions.

A practical process is to establish a long-run anchor using relative PPP or an estimated equilibrium real exchange rate, then assess cyclical deviations through monetary policy, growth, risk appetite and external financing. Fundamental forecasts should distinguish a forecast of the conditional mean from a stress range. A model that predicts long-run appreciation does not imply that the currency cannot depreciate sharply in the next quarter.

Technical forecasting

Technical analysis uses information contained in the exchange-rate series and market activity. Common tools include trend lines, moving averages, momentum, support and resistance, volatility measures and chart patterns. A simple moving average over n observations is:

An exponential moving average gives greater weight to recent observations:

Technical methods can be useful for timing, market sentiment and disciplined rule-based execution. However, apparent patterns may result from data mining, changing market regimes or transaction costs. A rule should be tested out-of-sample, with realistic bid–ask spreads, financing costs, slippage and limits to short selling.

Econometric forecasting

A baseline autoregressive model for the change in the log exchange rate is:

where c is a constant, φ_j are coefficients, p is the lag order and ε_t is an error term with conditional mean zero under the model. A vector autoregression (VAR) models exchange rates jointly with variables such as interest differentials, inflation, output and commodity prices. If variables are non-stationary but share a stable long-run relationship, a vector error-correction model (VECM) is appropriate:

Here x_t is a vector of variables, Π captures long-run adjustment and Γ_j captures short-run dynamics. Stationarity, cointegration rank, lag length, structural breaks and parameter stability must be investigated rather than assumed.

Forecast performance should be evaluated against a benchmark, often a random walk:

For a forecast \hat S_t+h, mean absolute error and root mean squared error are:

MAE is less sensitive to large errors; RMSE penalises them more heavily. Forecast comparisons should use rolling or expanding windows and, where appropriate, statistical tests such as the Diebold–Mariano framework. A statistically better forecast may still be economically inferior after transaction costs or hedging constraints.

Scenario-based forecasting

Because exchange-rate distributions are fat-tailed and policy regimes change, scenario analysis complements point forecasts. A disciplined scenario set might include:

Probabilities may be assigned when evidence supports them, but decision-makers should not confuse subjective probabilities with objective frequencies. Sensitivity analysis can show the effect on earnings, cash flow, debt covenants and investment value without pretending to know the exact future rate.

8. Integrating forecasts into managerial decisions

A robust forecasting system uses model diversity rather than unconditional faith in one model. One possible workflow is:

1. Define the exposure, currency quotation, horizon and decision date.
2. Collect consistent spot, forward, interest-rate, inflation, flow and market-position data.
3. Establish benchmarks: forward rate, random walk and PPP-based anchor.
4. Estimate fundamental and econometric models, testing for breaks and instability.
5. Add technical and market-implied information only after accounting for costs and risk premia.
6. Construct scenarios and quantify operational and financial consequences.
7. Make the decision using risk appetite, hedge policy and cash-flow certainty—not forecast direction alone.
8. Record the forecast, assumptions and outcome for ex-post validation.

Forwards reduce uncertainty but eliminate favourable upside; options provide asymmetric protection but require a premium; swaps can manage longer-term funding exposures; and operational hedges include currency matching, local sourcing, invoicing choices and geographic diversification. The correct instrument depends on the exposure’s timing, accounting treatment, liquidity and risk tolerance.

Illustrative case: an Indian importer of US medical equipment

Background. MedEquip India must pay US$10 million in six months. Its functional currency is the rupee, and its board requires a high degree of cash-flow certainty. The current spot rate is ₹83/US$. A six-month forward quote is ₹85/US$.

Problem. Management expects the rupee to strengthen because domestic inflation is falling, but a geopolitical shock could cause a sharp depreciation. It must decide whether to hedge the full payable.

Analysis. A spot-only payment would cost ₹830 million at today’s rate, but that amount is not guaranteed. A full forward hedge locks the rupee cost at:

The forward quote should not be treated as an unbiased forecast. It reflects the interest differential and possibly a risk premium. A forecast range might be ₹80–₹95 per dollar, generated from PPP, interest-rate analysis, market volatility and stress scenarios. At ₹95, the unhedged payment would be ₹950 million; at ₹80, it would be ₹800 million.

Approach. Suppose the treasury policy permits partial hedging. MedEquip hedges US$7 million forward and buys options on the remaining US$3 million, subject to a premium and approved counterparty limits. It also asks the supplier whether invoicing can be staggered and whether a portion of future export receipts can be matched against the payable.

Outcome. The company does not attempt to win a directional forecast. It protects the cash flow required by its budget while retaining some benefit if the rupee appreciates. The final outcome should be evaluated against the approved policy and risk-adjusted objectives, not merely against the ex-post spot rate.

Learning points. First, exposure measurement precedes forecasting. Second, the forward rate is a financing and hedging price, not necessarily an expected spot rate. Third, scenario analysis makes tail risk visible. Fourth, governance—limits, documentation, counterparty controls and independent valuation—is as important as model accuracy.

9. Limitations, governance and contemporary developments

Exchange-rate models face several structural limitations. Parameters may change after a regime shift; official statistics may be revised; inflation indices may not represent an international firm’s actual cost basket; and monetary policy may respond endogenously to the exchange rate. Correlation is not causation, and a model calibrated in a low-volatility period may fail during a crisis.

Forecast governance should include data lineage, version-controlled code, independent review, clear quotation conventions, model validation, stress testing and documentation of overrides. Derivative use requires attention to suitability, collateral, counterparty credit risk, legal enforceability, accounting rules and applicable financial-market regulation. Manipulating benchmark submissions, concealing losses or using derivatives outside delegated authority raises ethical and governance concerns.

Artificial intelligence and machine learning can process high-frequency prices, news, order flow and alternative data. Their advantages include nonlinear prediction and rapid updating. Their risks include look-ahead bias, unstable relationships, data leakage, opaque decisions, spurious correlations and poor performance under unprecedented regimes. AI should therefore be used with strict time-order validation, explainability appropriate to the decision, human oversight and controls against unauthorised trading. It is an analytical tool, not a substitute for economic reasoning or governance.

Contemporary practice increasingly combines market-implied information, central-bank communication, cross-asset indicators, climate and geopolitical scenarios, and probabilistic forecasts. The most defensible forecast is often a distribution or set of scenarios rather than a single number. For multinational firms, resilience comes from aligning forecasts with pricing clauses, natural hedges, liquidity buffers and flexible supply chains.

Chapter Summary

Exchange rates are relative prices whose interpretation depends on the quotation convention. Purchasing power parity links long-run exchange rates to relative price levels, while covered interest parity links forward rates to spot rates and interest differentials under no-arbitrage conditions. Uncovered interest parity concerns expected future spot rates and depends on risk-premium assumptions. The Fisher effect connects nominal rates, expected inflation and real returns.

Monetary models explain exchange rates through relative money supplies, incomes, interest rates and price adjustment; overshooting highlights sticky goods prices and rapid financial adjustment. Portfolio-balance models add imperfect asset substitutability, wealth, risk and capital flows. Forecasting methods include fundamental, technical, econometric and scenario-based approaches. Each has strengths and limitations, and all should be compared with simple benchmarks and evaluated out of sample.

For international managers, the objective is not to predict with certainty but to make decisions that remain robust across plausible exchange-rate outcomes. Exposure identification, appropriate hedging, scenario analysis, model governance and ethical control are essential complements to forecasting.

Key Terms

Exchange rate: Price of one currency expressed in another currency.

Spot rate: Exchange rate for near-immediate delivery.

Forward rate: Contracted exchange rate for delivery at a future date.

Appreciation: Increase in a currency’s value under a specified quotation convention.

Depreciation: Decrease in a currency’s value under a specified quotation convention.

Purchasing power parity: Theory linking exchange rates to relative price levels.

Relative PPP: Proposition that exchange-rate changes reflect inflation differentials.

Covered interest parity: No-arbitrage relation between spot, forward and interest rates.

Uncovered interest parity: Hypothesis linking expected exchange-rate change to interest differentials.

Fisher effect: Relationship between nominal interest rates, real rates and expected inflation.

Risk premium: Compensation required for bearing currency or other investment risk.

Real exchange rate: Relative price of goods adjusted for nominal exchange rates and price levels.

Monetary model: Model explaining exchange rates through relative money markets and prices.

Overshooting: Short-run exchange-rate movement beyond its long-run equilibrium level.

Portfolio-balance model: Asset-market model based on imperfect substitutability among assets.

Fundamental forecast: Forecast based on macroeconomic and financial determinants.

Technical forecast: Forecast based on historical price and market-activity patterns.

Cointegration: Stable long-run relationship among individually non-stationary variables.

Scenario analysis: Evaluation of outcomes under alternative coherent assumptions.

Natural hedge: Reduction of currency exposure by matching foreign-currency revenues and costs.

Review Questions

Short Answer Questions

1. What is the difference between a spot rate and a forward rate?
2. State the approximate relative PPP relationship and explain its variables.
3. Why is covered interest parity principally a pricing condition?
4. What assumption distinguishes UIP from CIP?
5. Define exchange-rate overshooting.

Long Answer Questions

1. Explain absolute and relative purchasing power parity, including the main reasons PPP may fail.
2. Derive the covered interest parity relationship for a domestic-currency-per-foreign-currency quotation and discuss its assumptions.
3. Compare the monetary and portfolio-balance models of exchange-rate determination.
4. Explain the Fisher effect and its relationship to UIP and PPP.
5. Evaluate the strengths and limitations of fundamental, technical and econometric forecasting.

Analytical / Application Questions

1. A country has 9% annual interest rates while its trading partner has 4%. If the spot rate is 50 domestic currency units per foreign currency, calculate the one-year covered-parity forward rate and explain why it need not equal the realised future spot rate.
2. Design a forecasting and hedging framework for an exporter receiving foreign currency revenues over the next twelve months. Specify data, benchmarks, scenarios, validation and governance controls.
3. An importer has a fixed foreign-currency payable but believes its domestic currency will appreciate. Analyse the trade-off between a full forward hedge, partial hedging and options, taking account of cash-flow certainty, upside participation and tail risk.

Multiple Choice Questions with Answers

Question 1

Under a domestic-currency-per-foreign-currency quotation, an increase in the spot exchange rate represents:

A. Domestic-currency appreciation

B. Domestic-currency depreciation

C. Foreign-currency depreciation

D. No currency movement

Correct Answer: B. Domestic-currency depreciation

Question 2

If domestic inflation is 8% and foreign inflation is 3%, relative PPP approximately predicts domestic-currency depreciation of:

A. 3%

B. 5%

C. 8%

D. 11%

Correct Answer: B. 5%

Question 3

Which instrument makes interest-rate parity a covered condition?

A. A forward exchange contract

B. A floating-price commodity contract

C. An uncovered foreign deposit

D. A domestic equity index

Correct Answer: A. A forward exchange contract

Question 4

With spot ₹90/US$, domestic interest 7% and foreign interest 4%, the one-year covered-parity forward rate is approximately:

A. ₹87.48/US$

B. ₹90.00/US$

C. ₹92.60/US$

D. ₹97.00/US$

Correct Answer: C. ₹92.60/US$

Question 5

The exact Fisher relationship is:

A. i = r + π^e in all cases

B. 1 + i = (1 + r)(1 + π^e)

C. i = r − π^e

D. 1 + r = (1 + i)(1 + π^e)

Correct Answer: B. 1 + i = (1 + r)(1 + π^e)

Question 6

Uncovered interest parity differs from covered interest parity because it uses:

A. A fixed exchange rate

B. Expected future spot exchange rates rather than a forward hedge

C. Only inflation data

D. Only trade-flow data

Correct Answer: B. Expected future spot exchange rates rather than a forward hedge

Question 7

In a simple monetary model, an increase in domestic money supply, holding other variables constant, tends to:

A. Appreciate the domestic currency

B. Depreciate the domestic currency

C. Eliminate inflation permanently

D. Leave the exchange rate necessarily unchanged

Correct Answer: B. Depreciate the domestic currency

Question 8

Which method is most directly designed to model a stable long-run relationship among non-stationary variables?

A. A VECM

B. A simple moving average

C. A random coin toss

D. A cross-sectional histogram

Correct Answer: A. A VECM

Question 9

Compared with MAE, RMSE:

A. Places greater weight on large forecast errors

B. Cannot be calculated for exchange rates

C. Always produces a lower value

D. Ignores forecast errors

Correct Answer: A. Places greater weight on large forecast errors

Question 10

Why should a forward rate not automatically be interpreted as an unbiased forecast of the future spot rate?

A. It may include a currency risk premium

B. It is always set by governments

C. It never reflects interest rates

D. It is denominated only in physical currency

Correct Answer: A. It may include a currency risk premium