Keynes Beyond Empire: Demand Management Under External Constraint

John Maynard Keynes developed his theory of aggregate demand in a specific institutional context: the interwar British economy, operating within the sterling monetary zone, with capital mobility constrained by convention and later by regulation, and with access to a colonial reserve system that provided external financing without the discipline that peripheral economies face when current account positions deteriorate. His later work at Bretton Woods in 1944 addressed the architecture of the post-war monetary order from the position of a power that, however diminished, remained within the core of the international monetary hierarchy. Both contexts shared a defining feature that Keynes did not theorise as such because it did not require theorising: the insulation of domestic demand management from external constraint.

That insulation is not available to most economies. The international monetary system is stratified, and fiscal policy transmits differently depending on a country’s position within that stratification. A finance ministry in Washington or Berlin that announces an expansionary fiscal package faces a different set of market responses than a finance ministry in Nairobi or Colombo that does the same. The difference is not merely one of credibility or institutional capacity. It is structural. It reflects the currency hierarchy, the balance-of-payments constraint and the import dependence of productive capacity that characterise peripheral economies. These conditions were not central to the Keynesian framework as originally developed. They are central to any application of that framework to the Global South.

This paper proceeds as follows. Section II establishes the balance-of-payments constraint as the binding external limit on fiscal expansion in import-dependent economies, drawing on Thirlwall (1979). Section III examines the empirical evidence on fiscal multipliers and their structural determinants, with particular reference to Blanchard and Leigh (2012) and Ilzetzki, Mendoza and Végh (2013). Section IV analyses currency hierarchy and the original sin problem as determinants of fiscal space. Sections V through IX examine the structural channels through which commodity dependence, financial architecture, industrial structure, climate investment and capital mobility alter the transmission of demand management. Section X proposes the framework of constrained Keynesianism. Section XI concludes.

The theoretical foundation for understanding fiscal policy in externally constrained economies is the balance-of-payments-constrained growth model formalised by Thirlwall (1979). The model’s central proposition — that an economy’s long-run growth rate is constrained by the ratio of its export growth rate to the income elasticity of demand for imports — follows from the simple accounting identity that external payments must balance over time without recourse to unsustainable reserve depletion or external borrowing. The growth rate consistent with balance-of-payments equilibrium is expressed as g = x/π, where x is the rate of growth of export volumes and π is the income elasticity of import demand.

The policy implication is direct and constraining. Fiscal expansion increases aggregate income and, in import-dependent economies, generates proportional increases in import demand. If the income elasticity of imports is high relative to the growth of foreign exchange earnings, the current account deteriorates as a mechanical consequence of stimulus. In Sub-Saharan Africa, empirical estimates of the income elasticity of imports range from approximately 1.2 to 2.0 (Bairam, 1993; Hussain, 1999; related IMF country studies), against an import-to-GDP ratio of approximately 33–35 per cent of GDP according to World Bank World Development Indicators 2023 data. These two figures in combination mean that a one percentage point increase in domestic income generates import demand growth of 1.2 to 2.0 times that income increment, applied to a base of one-third of GDP. The fiscal multiplier therefore competes with the import propensity at every step of the demand expansion.

The constraint does not operate as a sudden stop at a precisely measurable ceiling. It operates as a progressive tightening of financing conditions as the current account deteriorates, reserves deplete and exchange rate pressure accumulates. The IMF World Economic Outlook October 2024 database records a median current account balance of approximately negative 0.5 to negative 1.0 per cent of GDP across emerging market and developing economies, with approximately 12 to 15 Sub-Saharan African low-income countries running current account deficits in excess of 5 per cent of GDP in 2023. These are not positions that provide substantial margin for further demand-driven deterioration before adjustment mechanisms engage.

The empirical literature on fiscal multipliers has developed substantially since the 2008 financial crisis. Blanchard and Leigh (2012), examining forecast errors in IMF projections for advanced economies during the post-2008 consolidation period, found that actual fiscal multipliers during the crisis were substantially larger than the 0.5 assumed in standard IMF models, with estimates ranging from approximately 0.9 to 1.7. Their central finding — that multipliers exceeded unity in the specific conditions of the zero lower bound, depressed private demand and coordinated austerity — recalibrated the profession’s view of fiscal transmission in advanced economies under crisis conditions.

The conditions that produced high multipliers in the Blanchard and Leigh study do not generalise to open developing economies. Ilzetzki, Mendoza and Végh (2013), using a panel of 44 countries over four decades, found that the fiscal multiplier is close to zero in open economies with flexible exchange rates and is substantially higher in closed economies with fixed exchange rates. Their estimates for closed economies reach approximately 1.6 on impact, while open-economy multipliers are indistinguishable from zero in their central specification. The key transmission mechanism is the exchange rate: in open economies, fiscal expansion appreciates the real exchange rate (or, in developing country contexts, depreciates the currency and raises the cost of imported inputs), dampening the demand expansion and redirecting it toward imports rather than domestic value added.

The structural determinants of multiplier magnitude are therefore not merely technical parameters to be estimated and applied. They reflect the industrial and financial architecture of the economy. An economy that produces most of its consumption goods domestically, has a deep local-currency financial system and is not heavily dependent on imported capital goods will experience higher multipliers from a given fiscal impulse than an economy that is import-dependent across all three dimensions. Sub-Saharan Africa’s Economic Complexity Index of negative 1.02 (Harvard Growth Lab Atlas of Economic Complexity, 2022–23), its manufacturing share of GDP of 10.5 per cent against East Asia’s 24 per cent, and its household consumption import content of approximately 30–40 per cent (World Bank and OECD TiVA estimates) collectively indicate a structural multiplier environment closer to the Ilzetzki et al. open-economy finding of near zero than to the Blanchard-Leigh crisis-period finding of 0.9–1.7. This is not a policy failure. It is a structural condition that fiscal policy cannot correct by itself.

The asymmetry of fiscal policy effectiveness across the international monetary hierarchy was not a feature of Keynes’s analytical framework because Keynes was not writing for peripheral economies. The concept that formalises this asymmetry most precisely in the post-Keynesian literature is the “original sin” identified by Eichengreen and Hausmann (1999): the inability of most emerging economies to borrow internationally in their own currency. Eichengreen and Hausmann’s NBER Working Paper No. 7418, “Exchange Rates and Financial Fragility,” identified the structural condition in which peripheral economies must either borrow externally in foreign currency, exposing them to currency mismatch risk, or refrain from external borrowing, limiting investment and fiscal capacity. The concept was formalised further in Eichengreen, Hausmann and Panizza (2003), which estimated that the large majority of emerging economies — roughly 30 or more countries outside the major reserve-currency issuers — were unable to issue sovereign debt internationally in their own currency.

The consequences for fiscal policy are compound. First, external borrowing in foreign currency means that fiscal deficits carry exchange rate risk: a currency depreciation raises the domestic-currency cost of debt service without any corresponding improvement in the capacity to service it unless export revenues are also denominated in foreign currency. Second, the spread premium that peripheral sovereign borrowers pay relative to reserve-currency issuers reflects both credit risk and currency risk, and compounds over time into a structural fiscal disadvantage. JPMorgan EMBI data for 2023–24 indicate a spread premium of approximately 150 to 250 basis points between US and German sovereign borrowing costs and investment-grade emerging market sovereigns, with a further differential of approximately 300 to 500 basis points between investment-grade and sub-investment-grade EM borrowers. Against a pre-pandemic baseline of approximately 290 basis points on the EMBI Global Index and a current level of approximately 380 basis points, the cost of the external fiscal constraint is continuously repriced in sovereign spread markets. Third, the dollar accounts for 58.22 per cent of global allocated foreign exchange reserves (IMF COFER Q3 2024) and approximately 67 per cent of EM external debt is denominated in foreign currency (World Bank International Debt Report 2024), meaning that the currency hierarchy that conditions fiscal space is also the currency hierarchy that conditions reserve management and debt sustainability simultaneously.

The fiscal policy cycle in commodity-dependent economies is structurally inverted relative to the Keynesian prescription. Countercyclical fiscal policy requires the capacity to expand spending during downturns and consolidate during booms. In economies where government revenue is substantially derived from commodity royalties, export duties and profit-sharing arrangements with extractive industries, the revenue base expands precisely when commodity prices are high — that is, when the economy is already at or near capacity — and contracts precisely when prices fall, which is when countercyclical support is most needed. UNCTAD’s State of Commodity Dependence 2023 reports that 45 of 54 Sub-Saharan African countries and 101 countries globally are commodity-dependent, with commodities exceeding 60 per cent of merchandise exports. The World Bank Commodity Markets Outlook 2024 documents the amplitude of recent commodity price cycles: Brent crude ranged from approximately $19 per barrel in April 2020 to approximately $130 per barrel in mid-2022, a factor-of-seven swing. Copper ranged from approximately $4,600 to approximately $10,700 per metric tonne over the same period. Fiscal revenue systems exposed to these amplitudes cannot be designed as countercyclical instruments without explicit institutional separation between revenue flows and expenditure decisions — precisely the function of sovereign wealth funds and fiscal stabilisation mechanisms that are poorly developed across most of the Global South.

IMF Regional Economic Outlook analysis for Sub-Saharan Africa (2022) found that approximately 60 to 70 per cent of SSA countries exhibited pro-cyclical fiscal policy over the 2010–2020 period. Pro-cyclicality at this scale is not primarily a policy design failure. It is the predictable institutional outcome of fiscal systems in which revenue volatility is high, access to countercyclical financing is constrained by the currency hierarchy and sovereign spreads documented in Section IV, and political economy pressures to translate commodity booms into immediate expenditure increases are difficult to resist in the absence of credible rules-based fiscal frameworks. The SWF Institute estimates approximately 60 or more sovereign wealth funds globally with aggregate assets under management of approximately $10 to $11 trillion, but the Global South subset of this is substantially smaller and concentrated in a small number of commodity exporters. The majority of commodity-dependent economies have no meaningful stabilisation fund and face the full amplitude of commodity cycles directly in their fiscal balances.

The capacity of a domestic financial system to absorb countercyclical fiscal deficits without triggering instability depends on the depth of local-currency markets, the institutional investor base available to purchase sovereign paper, the degree of sovereign-bank nexus that amplifies fiscal stress into banking stress and the extent of foreign currency exposure within the banking system. Each of these dimensions is structurally constrained in most Sub-Saharan African economies.

IMF Global Financial Stability Report data for 2023–24 indicates that government securities constitute approximately 15 to 25 per cent of total assets in SSA commercial bank balance sheets. This concentration means that sovereign stress directly impairs bank capital, reducing the banking system’s capacity to extend credit to the private sector precisely when fiscal expansion is intended to support economic activity. The mechanism creates a direct conflict between the fiscal stabilisation objective and the financial stability objective that does not arise in the same form in economies with deeper non-bank financial systems. Approximately 8 to 12 SSA countries have domestic yield curves extending to 10 years or more (AfDB and IMF estimates), meaning that the majority of SSA sovereigns cannot reliably issue long-term domestic debt and must rely on shorter maturities that create rollover risk, or on external borrowing in foreign currency, reproducing the original sin constraint at the domestic level. Non-resident holdings of emerging market local currency government bonds average approximately 15 to 20 per cent of total issuance according to BIS Quarterly Review 2023 data, creating a sudden-stop dimension within local currency markets that was illustrated by the 2013 taper tantrum and the 2022 Fed tightening episode, both of which caused rapid and disruptive non-resident outflows from EM local currency bond markets.

The relationship between industrial structure and fiscal multiplier effectiveness is not incidental. If fiscal expansion generates demand that is satisfied primarily by imports — because domestic productive capacity is concentrated in low-complexity commodity sectors — then the fiscal impulse is transmitted to foreign suppliers rather than domestic workers and firms. The import content of household consumption in Sub-Saharan Africa is estimated at approximately 30 to 40 per cent of total consumption expenditure based on World Bank and OECD TiVA analysis. The import content of investment in low- and middle-income countries is estimated at 40 to 60 per cent of total investment, reflecting dependence on imported capital goods and intermediate inputs for which domestic substitutes do not yet exist. IRENA and IEA Africa Energy Outlook data for 2023–24 indicate that 70 to 90 per cent of renewable energy equipment installed in the Global South — particularly solar photovoltaic modules — is imported rather than locally manufactured. A climate investment programme in SSA that installs imported solar panels therefore leaks 70 to 90 cents of every dollar of investment to foreign equipment suppliers, with only 10 to 30 cents circulating through the domestic economy as multiplied income.

Export diversification improves the foreign exchange constraint by generating more stable and more diversified sources of foreign exchange earnings, reducing the sensitivity of the current account to single-commodity price cycles and widening the current account envelope within which fiscal expansion can occur before the balance-of-payments ceiling is reached. Sub-Saharan Africa’s Economic Complexity Index of negative 1.02, against East Asia’s positive 0.65 and the OECD average of positive 1.15 (Harvard Growth Lab, 2022–23), represents a structural multiplier environment in which the majority of additional demand generated by fiscal expansion is satisfied by imports before domestic productive capacity is engaged. The policy implication is that industrial policy aimed at domestic production of intermediate and capital goods, and export diversification toward higher value-added products, is simultaneously development policy and macroeconomic policy. It is not a supplement to demand management. It is the structural precondition for demand management to be effective.

Keynes was explicit about the relationship between capital mobility and domestic policy autonomy, in two distinct documents that are often conflated. His 1933 essay “National Self-Sufficiency,” published in the Yale Review (Vol. 22, No. 4, pp. 755–769), stated the position directly: “Ideas, knowledge, science, hospitality, travel — these are the things which should of their nature be international. But let goods be homespun whenever it is reasonably and conveniently possible; and, above all, let finance be primarily national.” His position at Bretton Woods in 1944 was more precise: he argued not for temporary or transitional capital controls but for permanent ones as a feature of the post-war monetary architecture, stating that the plan “accords to every member government the explicit right to control all capital movements” as a permanent arrangement rather than a transitional concession. This position was incorporated into the IMF’s original Articles of Agreement: Article VI, Section 3 explicitly permits members to “exercise such controls as are necessary to regulate international capital movements,” while prohibiting controls that restrict current account payments.

The IMF’s own position evolved away from Keynes over the following decades, as the Washington Consensus associated capital account liberalisation with market efficiency. The IMF’s formal reversal began with its Institutional View on Capital Flows, adopted by the Executive Board in 2012, which endorsed the use of capital flow management measures under specific conditions: when macroeconomic adjustment tools had been exhausted, when capital flows posed systemic risk and when controls were temporary and non-discriminatory. The IMF’s Integrated Policy Framework of 2022 went further, recognising the pre-emptive and integrated use of capital flow management alongside foreign exchange intervention as a legitimate first-line policy response rather than a measure of last resort. The IMF literature on sudden stop episodes — approximately 20 to 30 major episodes globally between 2000 and 2023 by BIS and IMF definitions — and on countries using capital flow management measures (over 40 countries maintaining some form of CFM in 2020–2023 according to the IMF Annual Report on Exchange Arrangements and Exchange Restrictions) demonstrates both the frequency of the problem and the growing acceptance of the instrument.

The country-level evidence supports the effectiveness of capital flow management under crisis conditions. Malaysia’s September 1998 capital controls, introduced in the context of the Asian financial crisis, were followed by an economic contraction of 7.4 per cent in 1998 and a recovery to 6.1 per cent growth in 1999, with controls lifted gradually thereafter. Iceland’s capital controls, introduced in 2008 in the context of the collapse of its banking system, contributed to a current account swing from a deficit of approximately 25 per cent of GDP in 2006 to surplus by 2010 (consistent with IMF Iceland country reports), with controls maintained until 2017. Neither case represents a cost-free intervention; both illustrate that capital flow management can provide the policy space for adjustment without the full force of sudden-stop transmission.

The energy transition introduces a specific form of the import content problem that is both large in scale and structurally distinct from conventional fiscal expansion. Renewable energy infrastructure — solar panels, wind turbines, battery storage, grid components — is manufactured primarily in China, the European Union and the United States. IRENA and IEA estimates for 2023–24 indicate that 70 to 90 per cent of renewable energy equipment installed in the Global South is imported, with solar photovoltaic module imports particularly dominant. A climate investment programme in an SSA economy that allocates, for example, $1 billion to renewable energy installation therefore generates approximately $700 to $900 million in demand for foreign-produced equipment — a direct current account cost — and approximately $100 to $300 million in domestic economic activity through installation labour, grid connection work and local services.

This is not an argument against climate investment. It is an argument for climate industrialisation alongside climate investment: the deliberate development of domestic manufacturing capacity for renewable energy components as a simultaneous objective with energy transition targets. The macroeconomic logic is identical to the industrial policy argument in Section VII. Green manufacturing capacity reduces the import content of climate investment, raises the domestic fiscal multiplier from climate spending and improves the current account position of the transition rather than deteriorating it. The parallel with Indonesia’s 2020 nickel ore export ban, which attracted approximately $30 billion in downstream processing investment within four years (BKPM data), is instructive: resource-based industrial policy that captures value-added domestically is simultaneously export diversification, current account strengthening and multiplier expansion.

The preceding analysis supports a revised framework for demand management in externally constrained economies that can be stated in five structural propositions. First, countercyclical fiscal policy must be calibrated to current account sustainability rather than to the output gap alone, because the balance-of-payments ceiling may bind before the output gap is closed. Second, the effective fiscal multiplier is endogenous to industrial structure, import content and financial depth, and policies to improve these structural parameters are therefore simultaneously macroeconomic policies. Third, local currency bond market development reduces the dependence of fiscal financing on external foreign currency borrowing, improving the currency composition of public debt and reducing exchange rate risk in the fiscal accounts. Fourth, capital flow management is a legitimate and documented instrument for preserving domestic monetary and fiscal policy autonomy under sudden stop conditions, consistent with the original Keynesian position at Bretton Woods and with the IMF’s own 2012 and 2022 institutional frameworks. Fifth, regional financial cooperation mechanisms — reserve pooling, regional payment systems, coordinated fiscal frameworks — can partially substitute for the liquidity backstops that reserve-currency-issuing central banks provide to their own economies and to a small number of permanent swap line counterparties.

Fiscal space, under this framework, is not a fixed endowment that governments either have or lack. It is a structural outcome that expands as export complexity increases, as local currency financial systems deepen, as import content of production falls and as regional cooperation widens the set of available financing instruments. The policy implication is that structural transformation — industrial policy, financial development, export diversification — is not an alternative to Keynesian demand management but its material precondition. Demand management without structural transformation risks balance-of-payments crises before full employment is reached. Structural transformation without demand management risks persistent underutilisation of the capacity that transformation creates.