Pakistan Macro Forecasting: Accounting, Scenarios and Market Transmission

Create an accounting-consistent starting point

1.1 Identities and units

Begin with nominal GDP, fiscal accounts, external accounts and monetary data on matching dates. The expenditure identity Y=C+I+G+X−M describes how output is measured; it does not prove that raising any component mechanically increases real output by the same amount. Separate nominal changes into quantity and price effects. A fiscal deficit is a period flow, while public debt is a stock affected by deficits, valuation changes and other adjustments. Mixing fiscal years, calendar years and monthly observations creates false relationships.

1.2 Stocks, flows and financing

Use a reconciliation sheet before fitting a regression. Express ratios using a consistent GDP vintage, distinguish central-government coverage from broader public-sector coverage and document whether figures are provisional. A financing identity constrains a scenario, but it does not tell you which investor will supply funds at which interest rate. A spreadsheet can balance through an unrealistic residual; that is a warning to investigate the economic mechanism rather than a successful forecast.

Step-by-step method

1. Define each series, unit, frequency and institutional coverage.
2. Reconcile opening stock, flows, valuation and closing stock.
3. Expose unexplained residuals instead of hiding them.

Model the balance of payments and financing gap

2.1 Current-account drivers

Build exports, imports, remittances, services and income flows from economic drivers. Import demand may depend on activity, energy prices and exchange rates, but elasticities and timing are uncertain. A current-account deficit is not identical to the trade deficit. Interest payments and profit remittances can matter even when merchandise trade improves. Match foreign-currency units and avoid adding a stock of reserves to a monthly trade flow without a time horizon.

2.2 Financial flows and reserve pressure

Use one declared balance-of-payments sign convention throughout. Project financial inflows, scheduled external payments and feasible reserve changes consistently, noting errors and omissions rather than treating them as dependable funding. An announced commitment is not the same as a disbursement. Test delayed financing, expensive energy and weaker receipts separately. Exchange-rate movement can affect import prices, debt valuation and company margins simultaneously, so a stronger external balance need not translate uniformly into equity gains.

Step-by-step method

1. Specify the sign convention and forecast horizon.
2. Separate committed, conditional and uncertain financing.
3. Calculate the residual gap and stress timing, not only annual totals.

Connect fiscal policy, inflation and monetary response

3.1 Multipliers are conditional estimates

A fiscal multiplier estimates an output response to a fiscal change under stated conditions. It varies with import leakage, spare capacity, financing, tax composition and monetary policy; it is not a permanent country constant. Separate spending announcements from implementation and distinguish one-off revenue measures from durable changes. Interest expense may respond with a lag to debt repricing, so a lower policy rate does not immediately reduce the entire budget interest bill.

3.2 Inflation and policy transmission

Construct inflation from clearly defined components and distinguish month-on-month momentum from year-on-year base effects. Energy adjustments, exchange-rate pass-through, weather and expectations may act at different lags. Do not describe Pakistan as operating a particular formal inflation-targeting regime merely because a model contains an inflation objective. Read current SBP policy explanations. A policy reaction function is an estimate of conditional behaviour, not a binding promise about the next decision.

Step-by-step method

1. State multiplier assumptions and alternative values.
2. Decompose inflation momentum, base effects and supply shocks.
3. Model policy and debt repricing with explicit lags.

Translate scenarios into companies and political-risk channels

4.1 Earnings and discount rates

Map macro variables to company revenues, input costs, working capital, debt and required returns. A currency depreciation may support an exporter’s receipts while increasing imported inputs or foreign debt. Banks face repricing, funding and credit effects; a rate change is not a universal sector signal. Index construction and concentration also matter: an aggregate GDP forecast cannot substitute for constituent-level earnings and valuation analysis. Avoid counting the same macro benefit in both earnings and an unjustified multiple expansion.

4.2 Political events without false precision

Political-risk quantification should begin with observable events and transmission mechanisms, such as implementation delay, financing disruption or regulatory change. Use conditional scenarios, named triggers and probability ranges where defensible. A subjective probability is not a measured historical frequency. Correlated shocks must be represented together when their causes overlap. Keep a narrative register explaining why an assumption changed so that revisions remain auditable rather than silently following market prices.

Step-by-step method

1. Map each shock to identifiable company cash-flow channels.
2. Separate earnings effects from discount-rate effects.
3. Specify political triggers and correlated downside cases.

Validate forecasts as they would have been used

5.1 Real-time vintages and benchmarks

Use rolling forecast origins with the releases available on each date. Revised historical GDP or inflation data can leak information into a backtest. Compare a complex model with simple alternatives, including an unchanged-value or transparent trend forecast. Report errors by horizon and regime, not only one favourable average. A narrow sample containing few crises cannot establish reliable tail probabilities, and tuning repeatedly to the same validation window weakens the evidence.

5.2 Scenario governance and revision

Maintain a forecast register with versioned data, assumptions, scenario probabilities and decision triggers. Assess point accuracy, interval coverage and whether the explanation of changes was credible. For investment use, include implementation costs and compare the market information already embedded in prices. A good economic forecast may still be a poor trade if anticipated or mistimed. Review errors to improve the mechanism, while preserving the original prediction rather than rewriting it after the outcome.

Step-by-step method

1. Freeze data and assumptions at each forecast origin.
2. Compare out-of-sample errors with simple benchmarks.
3. Publish revisions, uncertainty and the investment transmission limits.

From economic accounts to a testable forecast

Reference table

Glossary of key terms

Accounting identity

A relationship that must reconcile by definition.

Financing gap

Needs not covered by identified feasible funding.

Fiscal multiplier

Estimated output response to a fiscal change under stated conditions.

Forecast vintage

Data and assumptions available at a specific forecast date.

Pass-through

Transmission of one price or policy change into another variable.

Scenario

A conditional internally coherent set of outcomes.

Frequently asked questions

Common pitfalls