DIALOGUE
N°2- Décembre 1994

DEVALUATION : THE EXPERIENCE OF ANGLOPHONE AFRICA
Paul Collier, Director, Centre for the Study of African Economies, Oxford

Many of the countries of Anglophone Africa have undertaken devaluations of an even greater scale than that of the CFA Franc. They were, however, rather different. The devaluations of Anglophone Africa were primarily or entirely trade-liberalising, whereas that of the Franc Zone is primarily payments-improving. This can be seen most clearly in the case of the Nigerian devaluation of 1986. Prior to the devaluation the massive demand for foreign exchange was rationed to the available supply from export earnings through a series of trade restrictions, sometimes taking the form of import bans and more usually through limited access to foreign exchange licences. The Nigerian economy was thus already living within its means, albeit inefficiently. Those imports which were permitted into the economy sold at market-clearing prices, the difference between this and the cif import price at the official exchange rate accruing as rents to the distribution system and those controlling the allocation of permits. As a result, the devaluation neither reduced imports nor raised prices. In the year following a fourfold devaluation inflation was only around 5%. This has been the general pattern in Anglophone Africa, with devaluations enabling the removal of highly inefficient ways of rationing imports. Since the devaluations were usually part of policy packages supported by programme aid, the balance of payments often "deteriorated" following devaluation, though since this was intrinsic to the utilisation of the aid it should not be seen as a policy problem.

One important difference between payments-improving and trade-liberalising devaluations is their inflationary consequences. In the circumstances of the Franc zone some inflation was desirable. It reduced the real value of government domestic indebtedness and reduced real wages in those sectors where nominal wages were sticky downwards. Although trade-liberalising devaluations are not inflationary, Anglophone Africa had no need of devaluation-induced inflation. The underlying inflation rate had already been so high due to the absence of fiscal restraints that governments and other public entities had effectively defaulted on their domestic liabilities, and any nominal rigidities in the labour market had either been eroded or resulted in real wages below equilibrium levels. For example, in Uganda the post office savings bank had kept much of its assets in Sterling while its liabilities were denominated in Ugandan shillings at low or zero nominal interest rates. Over the course of thirty years the liabilities declined to less than 5% of the value of the assets. Inflation also eroded real wages in the public sector to levels at which it was not possible to live, in the process changing the nature of the public payroll from the purchase of labour services to a transfer scheme. Thus, Anglophone Africa needed devaluation neither for balance of payments improvement, nor to change the real value of wage and debt contracts.

One aspect which the Anglophone devaluations had in common with that of the CFA Franc is their fiscal impact. This depends upon whether the governments are net buyers or sellers of foreign exchange. Since most governments either taxed trade heavily or received a net aid inflow which they sold domestically, the fiscal impact would normally be positive, permitting increased government expenditure and reduced deficits. For example, over a decade of devaluations in Ghana the government was able to triple revenue as a share of GDP, while in Nigeria the 1986 devaluation was so revenue-enhancing that almost the entire fall in income as a result of the crash in the world oil price was shifted to the private sector.

Nigerian Income (in million 1984 Naira)
  Public Private
1985 23,574 78,476
1986 22,345 57,458
1987 20,756 59,382

Source : D.L.Bevan, P.Collier and J.W.Gunning, Nigeria 1970-90, ICEG, San Francisco, 1993.

Have the trade-liberalising devaluations of Anglophone Africa worked in their primary objective of improved resource allocation? Such improvements come partly from increased intra-sectoral allocative efficiency arising from a better allocation of imports, and partly from increased inter-sectoral allocative efficiency arising from the shift of resources into the export sector (primarily from the import-competing sector). The latter effect can be measured by the post-devaluation performance of exports, but the former is more diffuse, implying a generalised increase in growth. That the Nigerian economy could grow more rapidly in the five years after the 1986 devaluation than it had done during the oil boom of the 1970s, despite low investment, is some indication of how detrimental foreign exchange rationing has been to overall growth. Further evidence comes from a recent econometric study of the consequences of foreign exchange rationing as measured by the parallel market premium (Easterly and Levine, 1994, World Bank mimeo). Comparing Africa and East Asia they found that the African annual growth rate would have been higher by 0.64% had the premium been reduced to the East Asian level. The export performance of those Anglophone African countries which devalued has been only moderately encouraging, although Ghana, the country with the longest sustained devaluation experience has had export growth of around 8% p.a. post-1987. The failure of exports to grow more rapidly reflects a reluctance to commit investment to the export sector : labour moves because such relocation is reversible, whereas capital does not move because it would be irreversible.

The weakness of the export response to Anglophone African devaluations is at root a reflection of the inability of governments to pre-commit to the maintenance of policies other than through the fragile device of donor conditionality. The private sector, both domestically and internationally, sees trade liberalisation as less than fully credible. Anglophone Africa lacks an institution equivalent to the Franc Zone which, through a system of reciprocal threats has to a degree pooled sovereignty to create some restraints upon the policies of national governments. This has produced very wide swings in the floating exchange rates of Anglophone Africa, reflecting changes in the assessment of prospective policies. For example, the Kenyan shilling, the Ugandan shilling and the Zambian Kwacha have all appreciated very sharply in the past year as confidence has recovered from very low levels. Very wide fluctuations in exchange rates both reflect underlying uncertainty and are directly a considerable source of risk for the export sector. For several years prior to the devaluation of the CFA Franc, Africa was thus paradoxically divided into a Francophone group which was credibly locked in to an inappropriate exchange, and an Anglophone group which had appropriate exchange rates which were incredible because of the absence of lock-in institutions. The denouement of liberalisation without lock-in in Anglophone africa has been the collapse back to foreign exchange rationing in Nigeria, a demonstration of unrestrained economic illiteracy which will have set back confidence in that country for many years.

A further econometric result of Easterly and Levine is that there are powerful growth effects in Africa from neighbours. This implies that Nigeria suffered in recent years from the poor performance of the Franc Zone whereas currently the Franc Zone is suffering from the poor performance of Nigeria. The lack of supra-national institutions promoting policy coordination and limiting policy error has exacted, and continues to exact, an avoidable cost which Africa cannot afford to pay.


DIAL

- DIALOGUE N°2: Sommaire -