Developing indicators to diagnose Dutch Disease

Index

List of Figures

List of Tables

List of Abbreviations

1 Problem definition and objectives

2 Conceptual demarcation and analytical framework
2.1 Conceptual embedding of the DD
2.2 Role of causes and symptoms of the DD
2.3 Distinction of real and monetary type of the DD
2.4 Distinction of static and dynamic analysis of the DD
2.5 Introduction to resource movement effect and spending effect

3 The real type of the DD
3.1 The Core Model: effects of the boom when labour is mobile
3.1.1 The pre-boom equilibrium
3.1.2 Resource movement effect with mobile labour
3.1.3 Spending effect with mobile labour
3.1.4 Combining resource movement effect and spending effect
3.2 Differentiating favourable shocks
3.2.1 Non-neutral technological progress
3.2.2 Resource boom
3.2.3 Price boom of an already produced traded good
3.2.4 Exogenous inflows from abroad
3.3 Policy of absorption emphasis
3.3.1 Public consumption
3.3.2 Public domestic investment
3.3.3 Transfers to the private sector
3.4 Extensions of the Core Model
3.4.1 Labour market disequilibrium regimes
3.4.1.1 Classical unemployment
3.4.1.2 Repressed inflation – labour shortage
3.4.2 Price boom and domestic use of the booming good
3.4.3 Decomposition of the Lagging Sector into export and import sector
3.4.4 Mobile capital between the Lagging and the Non-traded Sector
3.4.5 Mobile capital within a diversified Lagging Sector
3.4.6 Mobile capital between the Lagging, Non-traded, and Booming Sector
3.4.7 Sector-specific capital but international capital mobility
3.5 Interim recapitulation

4 Monetary type of the DD
4.1 Monetary adjustments to a boom in two exchange rate regimes
4.1.1 Floating exchange rate after a boom
4.1.2 Fixed exchange rate after a boom
4.2 Monetary consequences of international capital mobility
4.2.1 Interest opportunities and resource depletion
4.2.2 Endogenous impact of a boom on the real interest rate
4.3 Monetary policy options
4.3.1 Nominal exchange rate policy
4.3.1.1 Devaluation in a fixed regime
4.3.1.2 Revaluation in a fixed regime
4.3.2 Money supply policy
4.3.2.1 Fiscal policy as a driver of money supply
4.3.2.2 Fiscal policy preventing an increase in money supply
4.3.2.3 Keynesian unemployment due to money supply rigidity
4.4 Interim conclusion

5 Plain dynamics of the DD
5.1 Dynamics of the Booming Sector
5.1.1 Time path of consumption spending
5.1.2 Time path of domestic investment
5.2 Dynamics of exogenous favourable shocks
5.2.1 Foreign aid and remittances
5.2.2 Foreign direct investment
5.2.3 Foreign debt
5.3 Dynamics of the Lagging Sector
5.3.1 Intertemporal adjustment with no market imperfections
5.3.2 Intertemporal adjustment with learning-by-doing externalities
5.4 Policies of supporting the Lagging Sector
5.4.1 Subsidisation policy
5.4.2 Trade protection
5.4.3 Exchange rate protection
5.5 Interim conclusion

6 Definitory conclusions for the DD

7 Measurement approaches to the DD
7.1 Measuring the relative prices
7.1.1 Empirical complications of the real exchange rate
7.1.2 Empirical advantage of the real effective exchange rate
7.2 Differentiating the disease from natural adjustments
7.3 Distortions of the DD
7.4 Role of the real (effective) exchange rate
7.5 Role of sectoral changes
7.6 DD index of Chenery and Syrquin

8 Empirical research
8.1 Data sources
8.2 Determination of the election criteria of a DD suspicion
8.2.1 Achievement of comparability
8.2.2 Reduction to an adequate scale
8.2.3 Research order of causes and symptoms
8.2.4 Determining the causes
8.3 Determination of DD indicators
8.3.1 Weighting of changes in exchange rate & prices
8.3.2 Weighting of sectoral shifts
8.3.3 Weighting of inflow management
8.3.4 Weighting of national economic growth
8.4 Election of the main DD period
8.5 Indexation of the DD intensity
8.6 Evaluation
8.6.1 Overview of cases at sufficient DD intensity levels
8.6.2 Comparing elected DD cases with the literature
8.6.3 Possible weak points of the study
8.6.4 Ratios of probable DD cases in total cases

Appendix

References

Eidesstattliche Erklärung

List of Figures

Fig. 1 (left): The commodity market for non-tradeables

Fig. 2 (right): The pre-boom equilibrium in the commodity market.

Fig. 3 (left): Resource movement effect (a to br) and subsequent new equilibrium (b to c) in the market for N

Fig. 4 (right): Resource movement effect (a to br) and subsequent new equilibrium in the commodity market for N and B & L

Fig. 5 (left): Consumption spending effect (a to e) in the market for N

Fig. 6 (right): Consumption spending effect (a to e) in the commodity market for N and B & L

Fig. 7 (left): Resource movement & consumption spending effect (a to f) in the market for N

Fig. 8 (right): Resource movement & consumption spending effect (a to f) in the commodity market for N and B & L

Fig. 9: Direct (L to L’) and indirect (L’ to L’’) de-industrialisation shown in the labour market.

Fig. 10: Consumption spending effect (a to i) in case of a price induced windfall

Fig. 11: Consumption spending effect (a to k) in case of a price induced windfall partly accruing to the government

Fig. 12: Production and absorption possibilities change when the windfall is fully spent on imported investment goods

Fig. 13 (left): Full investment into the Non-traded Sector without productivity enhancement of the traded sector

Fig. 14 (right): Full investment into the Non-traded Sector with non-traded goods turning into tradeables and productivity enhancement of the traded sector

Fig. 15: Investment leading to real appreciation and a Lagging Sector

Fig. 16 (left): Disequilibrium regimes with wage and price rigidities

Fig. 17 (right): Effects of a boom with wage and price rigidities

Fig. 18: Effects of the boom when capital is mobile between L and N

Fig. 19: Real and monetary effects of a boom under fixed and floating exchange rates with flexible wages and prices

Fig. 20: Boom-induced exchange rate overshooting and transitional unemployment due to sticky prices of N

Fig. 21: Monetary adjustment in DD and effect of devaluation

List of Tables

Table 1: Comparison of standardisation methods

Table 2: Possible causes of the DD

Table 3: Choice, weighting and DD direction of possible indicators

Table 4: Country cases per year with DD intensities >= 0.4

Table 5: Country cases per country with DD intensites >= 0.4

Table 6: Example of the calculated relative changes

Table 7: Percentage rank of relative changes

Table 8: Average of relative changes

Table 9: Percentage rank of the average of relative changes

Table 10: Average of each year’s rank of the relative change

Table 11: Averaging the averages as the final intensity for the indexation

Table 12: Three index value levels and four data point levels

List of Abbreviations

illustration not visible in this excerpt

1 Problem definition and objectives

The term Dutch Disease (abbreviated to DD throughout the paper), introduced in 1977, refers to the adverse effects on Dutch manufacturing of the natural gas discoveries of the 1960s.^[1] The crucial sub-period from 1974 to 1979 after the oil price shock in 1973 / 1974 was then marked by a consumption driven booming government sector in the context of a European stagnation, and though this process in itself did not bear a disease character, the strong real appreciation due to an overvaluation of the Dutch Guilder and an inflexible labour market were at least the clearest possible signs of a disease.^[2] It is actually doubted that the DD is Dutch, but the existence of the DD as a general phenomenon is widely accepted in the literature.^[3]

The fascination of the DD arises from its paradoxical nature that something intuitively good develops a dark side.

There is a good side of every boom, its initial impact is beneficial and amounts to Pareto-improvement for the economy as a whole.^[4] This implies a rise in real living standards due to higher levels of public and private consumption and higher levels of investment (and savings).^[5] Windfalls, the linked fiscal revenues, and easier domestic and foreign borrowing can finance core public goods.^[6] Especially regarding developing countries, windfalls principally allow breaking out of the poverty trap: poverty – lack of public finance – lack of public goods – lack of private investment – poverty.^[7]

Approaching the dark side, a favourable shock like a discovery of oil is a mixed blessing to developing countries, and research does not show a clear outperformance of the oil nations as a whole.^[8] The DD can be formulated as a provoking dilemma: “Enjoy boom revenues to boost economic development while those revenues in fact turn out to be responsible for the economic stagnation through the deterioration of the tradeable sectors.”^[9]

The provoking nature of the DD runs the risk of being misused in the media in the longing for exciting economical statements, and a healthy mistrust should always accompany an alleged DD even in the literature, as the real problem can lie somewhere else in the dynamic economy. For example, in many countries, the downward trend in the share of manufacturing in national output is dominated by other reasons than a favourable shock.^[10] In general, policy choices can exacerbate or mitigate the DD, so that there is a direct connection, but the majority of influences rather works independently from the DD and is, if at all, indirectly linked with it.

A disease typically calls for a cure. Political issues of the DD cover the distribution of the gains, transitional assistance to declining sectors and the appropriate response to various market failures which may impede the smooth adjustment of the economy to its new equilibrium.^[11] This paper searches mainly for the pure disease character but does not exclude the curing counterparts or exacerbating distortions that support a more realistic view on the disease’s extent.

The concluding course of investigation is as follows.

Beginning with the structuring of its theoretical background, the DD is then split into a real type of the DD, with a focus on structural changes in a rather static framework, and a monetary type of the DD that focuses on monetary adjustment problems associated with a boom or something similar that stresses more the problems of the adjustment phase and is thus rather dynamic.

The subsequent section’s pure dynamics of the DD presents the central idea of the DD and integrates both the real and monetary type, though stressing the real type of the DD as its core.^[12]

The main extensions and distortions of models and results are explained where adequate in these three main sections, and the remaining distortions that are only indirectly connected with the DD will be explained in a separate section.

The theoretical analysis is ended with the main definitory conclusions that form the DD hypothesis which is not a definition in the classic sense but a conglomerate of definitions. A short definition of the DD would probably miss the point, and this is the reason why it is not forestalled in this section. These theoretical conclusions shall also substitute a conclusion at the end of the paper.

The comprehensive theoretical analysis is complemented by the measurement methods in the literature and eventually followed by the empirical research as a practical implementation of the theoretical part. The research objective will not be to prove the existence of the DD which would require thorough qualitative country studies. The objective is to develop a quantitative method of diagnosing the DD in a way of determining its intensity with simple descriptive statistics, and this on a world-wide scale. The DD candidates with comparatively high intensities of the main indicators in neighboured years can finally be expected to have it. The research approach is thus focused on electing the potential DD candidates on an aggregate level and on setting the possible DD cases in relation to each other, allowing for a final indexation of the DD intensity and two world-wide overviews. A sample of this overview is tested on its accordance with the results from the literature. And though there are still weak points left and discussed, the question of how to diagnose the DD is solved to the main extent. Lastly, there is some information about the probable main disease cases in percent of total cases.

2 Conceptual demarcation and analytical framework

2.1 Conceptual embedding of the DD

The DD is associated with the "Booming sector" theory that stresses the impact of a Booming Sector on resource allocation in the rest of the economy.^[13] Apart from this Booming Sector theory, the DD can also emanate from exogenous inflows which do not necessarily create a Booming Sector, but all favourable shocks have in common that there is a Lagging Sector in the end.

The DD seems to be nothing new in theory. The adversely affected sectors and relative price changes which accompany a boom are an old subject both in the historical and the analytical sense.^[14] For example, the idea that labour-saving technical progress can have an adverse effect on labour even though it is potentially Pareto-improving is an old one.^[15] The issue also arose in the 1950s when the United States gave food aid to many countries while possibly harming those country’s food production.^[16] The special feature of the DD is now the view on factor incomes changing due to effects on sectors.^[17]

Many historical episodes of sectoral booms with adverse general equilibrium effects on other sectors could be looked at in terms of DD and Booming Sector concepts.^[18]

2.2 Role of causes and symptoms of the DD

Looking at the DD as a disease in the literal sense, there must be causes and symptoms that can be diagnosed. But it might happen that the causes of the DD occur without any symptom, and the symptoms occur without any causes. On the one hand, there may exist causes which are not that apparent, and symptoms which are dominated by counter-developments. On the other hand, it is possible as well that there is no decisive link between alleged causes and symptoms, and then, there is no possibility of diagnosing a disease at all. In fact, separating causes from symptoms would allow just speaking of international trade phenomena, but not of a disease. Most economies have a comparative advantage that has lead to a focus on certain sectors, and most economies change their sectoral structure in time due to international influences.

Though the Booming Sector theory is linked to a discovery of resources or a resource price boom or a quick technological process, there are other favourable shocks, and it is obvious that the causes are not essential in this theory. A major part of causes will be explained in section 3.2, but the list is most probably not exhaustive.^[19]

The main DD symptom is the existence of a lagging non-booming tradeable sector.^[20] This Lagging Sector can be producing both non-boom exportables and importables, and in the frequent DD case of a resource-based boom it can cover all non-raw-material branches that produce tradeable outputs.^[21]

Since the causes alone obviously do not make the alleged disease, the following theoretical DD analysis starts by focusing on the symptoms and later describes the causes and their implications in detail, while the empirical research will start with the causes, and the won data will afterwards be checked for symptoms.

2.3 Distinction of real and monetary type of the DD

There are two main types of the DD.

The real type of the DD is motivated by real trade theory analysis where sectoral effects in the real economy are examined.^[22] It is the core of the DD, since the primary impact of any favourable shock falls on the level of real income and on the intersectoral allocation of factors of production.^[23] Hence, it is natural to begin the analysis with a real model which abstracts from monetary considerations.^[24]

The monetary side of the DD, on the other hand, essentially comes from short-run exchange rate responses combined with price stickiness, and it is not an unanimous part of the DD, as it could be ascribed to a plain stabilization problem, and nevertheless, these effects seem to play an at least occasional role for the DD.^[25]

2.4 Distinction of static and dynamic analysis of the DD

The necessary link between causes and symptoms makes diagnosing the DD so difficult in a real complex economy with so many different impacts. As a result, a definition of the DD should abstract from these complexities by using simple models which may be found as a whole aggregate hidden in a complex environment. Diagnosing the DD is then the try to systematize a systemic process that is embedded into both very complex (i.e. static) and even more complicated (i.e. dynamic) environment. The process, the DD itself, is of an intrinsically dynamic nature since it necessarily has a time component embedded and becomes a true disease only when integrating dynamic considerations.^[26]

Fortunately, the theoretical approach of the DD can be simplified by separating the static view on the DD from the dynamic view.

Within the static view, the DD can be put into a real dimension and a monetary dimension, while the intertemporal dimension is ignored by taking the occurrences of a certain period as one aggregate event.

The dynamic view, though, bears a much more systemic character, which is barely controllable in a research context. Thus, the dynamic view has only the theoretical purpose of hypothesizing the DD, and this, paradoxically, just because it tries to be more realistic: the DD would not be a disease without the dynamics which go far beyond the normal adjustment processes.

This paper’s theoretical analysis of the DD is primarily concerned with the short- and medium-run effects of asymmetric growth on resource allocation and income distribution in both static and dynamic contexts, while the longer-run issues are reduced to the core.^[27] Thus, this paper will focus on the ‘upside’ of a favourable shock. The downside is neglected since it incurs too many distortions from long-run dynamics of an economy that decrease the explanatory power, and large scale empirical research would be too complicated.

2.5 Introduction to resource movement effect and spending effect

A resource boom or something equivalent affects the economy and its sectoral structure in two ways, namely by a resource movement effect and a spending effect, together forming a new general equilibrium.^[28] The effects can be applied both in a static and dynamic context. In the simplest static model, the resource movement effect draws factors of production out of other activities and into the Booming Sector, and the spending effect draws factors of production out of activities producing non-booming traded commodities (to be substituted by imports) and into nontraded sectors.^[29]

Only if the Booming Sector is not an enclave but requires a significant input of productive factors which must be bid away from other sectors in the economy, the resource movement effect applies.^[30] Thus, the resource movement effect can probably be ignored in the case of oil, since the use by the oil industry of domestic inputs, including labour, would be low in relation to their use by the rest of the economy.^[31] Natural resources in general are likely to be extracted with relatively few resources without affecting the domestic factor markets, especially if there is a discovery of a completely new resource.^[32] Apart from this, the resource movement effect is often buffered due to a usually given pool of unemployed workers.

In the analyses of this paper, both the resource movement and the spending effect are considered, though the spending effect matters most when diagnosing the DD as it is generally expected to dominate at a given considerable favourable shock.

3 The real type of the DD

The following analysis considers only the aspects of the DD that concern the allocation of real resources in a static framework.^[33] The static analysis can only show the necessary background of the DD, but the actual disease character lies in its dynamics. Nevertheless, processes that are analysed in the static framework are principally necessary for exerting or influencing the DD. The literature focuses on the static effects of a boom, the majority concentrating on the short- and medium-run effects of asymmetric growth which takes place at an exogenous rate and thus takes the time path of extraction and revenues as given.^[34]

“Static” indicates that the timeline of changes is ignored, everything is assumed to happen at the same time, though this only approximates reality.^[35] It also indicates that the balance of payments on current account (strictly the balance of trade) stays constant.^[36]

The focus on the structural aspects of a boom leaves aside monetary considerations with nominal variables.^[37] The now used simple static framework will examine a boom’s consequences on the functional distribution of income, relative prices, the size and profitability of the exposed traded sector and the level of unemployment.^[38]

There are three categories of models to explain the new general equilibrium. The first, which also represents the Core Model, shows the effects of the boom when labour is the only mobile factor. This model is probably most appropriate to the relatively short time horizon which is so significant for the DD research due to the many possible longer-run distortions.^[39]

The second category adds capital as another mobile factor, but keeps its mobility between the non-booming Traded and the Non-traded Sector. The third category assumes perfect mobility of capital and labour between the booming and both non-booming Traded and the Non-traded Sector.

The latter two models are explained in the section dealing with the extensions of the Core Model and will refer to its main assumptions.

There are many causes of the DD, but for now it suffices to presume the simple case that there is a costless once-for-all ‘Hicks-neutral’ improvement in technology; and though other causes of the boom have partly different effects, they all can be reconstructed on the base of the same model and the consequences are not crucially different.^[40] An example for a non-extractive Booming Sector is a sudden country-specific progress in know-how or technology, for example due to a displacement of older industry by technological advanced activities.^[41]

As was said in section 2.2, the causes are not the main characteristic of the DD. The various causes and their different consequences will thus be explained after the Core Model in Chapter 3.2.

3.1 The Core Model: effects of the boom when labour is mobile

The so-called 'specific factor' model assumes that each sector uses a single specific factor as well as drawing on a pool of intersectorally mobile labour.^[42]

The specific factor is assumed to be capital, but other options are possible, like some skill categories of labour as immobile (especially in the short run), and the specific factor in the energy sector can be thought of as including natural resources as well as specific capital.^[43]

There are three sectors. Assuming a small open economy, the Booming Sector “B” and the Lagging Sector “L” produce a composite traded good (Xb and Xl) whose price is fixed exogenously by world prices, while the Non-traded Sector “N” produces a single non-traded good Xn whose price is determined endogenously by the interaction of domestic supply and demand.^[44]

There are no distortions in commodity or factor markets.

In the commodity market, the Booming Sector does not deliver intermediate inputs for other sectors, but produces direct consumption commodities.^[45] All goods are normal goods, so that a positive income shock results in extra spending on both tradeables and non-tradeables.^[46] Terms of trade are given, i.e., the relative price of the two traded goods Xb and Xl does not change.^[47]

In the factor market, real wages are perfectly flexible, ensuring that full employment is maintained at all times.^[48] There is no leisure-consumption trade-off, so that boom revenues and other inflows resemble an increase in income of households.^[49] Hence the boom must raise potential national welfare, and the analysis can focus on the distribution of the gains among different factors in this framework.^[50]

3.1.1 The pre-boom equilibrium

illustration not visible in this excerpt

Fig. 1 (left): The commodity market for non-tradeables

Source: Cordon / Neary (1984): 361

illustration not visible in this excerpt

Fig. 2 (right): The pre-boom equilibrium in the commodity market.

Source: Humphreys / Sandbu (2007): 182, own thoughts

The supply curve S0 in Fig. 1 is derived from the transformation curve ‘NT’ between the non-tradeables Xn and the two tradeables Xt, as depicted in Fig. 2: infinite particular price ratios and thus infinite budget lines tangent to the transformation curve determine the supply curve (as adumbrated by the three budget lines on transformation curve NT with the equilibrium in point a in both figures).^[51] The demand curve D0 in Fig. 1 is not independent of the supply curve, since expenditure equals income by assumption and thus depends on output (the point where the budget line is tangent to the transformation curve in Fig. 2), and the demand curve D0 is determined by the chosen consumption points at the particular price ratio associated with the different output and income levels.^[52] As a result, both demand and supply curve account for each other. Mind that absorption and production possibilities frontier as the transformation line in Fig. 2 and simultaneously demand and supply curve in Fig. 1 lose more and more dependence from each other the more income diverges from output.

3.1.2 Resource movement effect with mobile labour

illustration not visible in this excerpt

Fig. 3 (left): Resource movement effect (a to br) and subsequent new equilibrium (b to c) in the market for N

Source: Corden (1984): 360, own thoughts

illustration not visible in this excerpt

Fig. 4 (right): Resource movement effect (a to br) and subsequent new equilibrium in the commodity market for N and B & L

Source: Corden / Neary (1982): 829, own thoughts

As a result of the boom, profitability and thus the marginal product of labour in B rises, so that, at a constant wage in terms of tradeables, the demand for labour in B increases, inducing a movement of labour out of L and out of N into B.^[53] The resource movement effect stands for this labour migration from the non-booming sectors into the Booming Sector at a constant real exchange rate, shifting the supply curve leftwards from S0 to S1 in Fig. 3.^[54]

In the Salter diagram of Fig. 4, the boom shifts out the transformation curve from NT to NT’, and this shift of the production possibility frontier is asymmetrical, since the maximum output increases only for traded goods.^[55]

On the demand side, consumption of N does not change in the new threefold intersection between the vertical income-consumption curve E0r, the new budget line at the initial exchange rate (= slope of the tangent at the production possibilities curve in Fig. 4), and the indifference line which depicts the summarised aggregate demand.^[56] The vertical line depicts the movement of the demand curve at the initial real exchange rate as the production possibilities rise, so that the indifference curves do not shift aside and thus any spending effect is excluded; demand for N does not react to the higher production, since the income-elasticity of demand for N is zero for the resource movement effect.^[57] This can be seen more intuitively in Fig. 3, where the demand curve of N (D0) simply does not shift.

But there is a decrease in the output of N at a constant real exchange rate, and the consecutive excess demand for N in point ar (which is equivalent to point a in Fig. 3) – shown in Fig. 3 and Fig. 4 by the brackets – can only be outbalanced by a real appreciation, as the arrow indicates.^[58]

The equilibrium in point c in Fig. 3 is reached with a higher price of N, so that PN / PL grows.^[59] This real appreciation (i.e., a rise in the relative price of non-traded goods) following the resource movement effect makes production of the nontradeables relatively more profitable than the production of the Lagging Sector’s tradeables, switching demand away from services and dampening the fall in the N Sector’s output by recruiting labour from L and B.^[60] The now assumed relative price change thus allows for a change in demand for N, and for infinite relative price changes the income-consumption line thus changes to E1a, intersecting with a new and not depicted budget line and the according indifference curve in point c. Importantly, this process after the resource movement effect is still abstract from the spending effect of the next section.

3.1.3 Spending effect with mobile labour

The spending effect should be modelled in isolation, as was the resource movement effect, since both together would distort each other. This approach is also especially useful, since the spending effect can happen independently and is the main process of the DD.

For a simple beginning, the higher income due to the boom is assumed to be spent only on consumption goods; the spending on investment goods will be explained in section 3.3.2, showing that proper investments can mitigate the DD.^[61]

The cause of the boom is still assumed to be a technological improvement, but it does not induce an increased B Sector’s demand for labour this time. To exclude this resource movement effect, the transformation curve must shift up vertically with bs vertically above a, as Fig. 6 shows, excluding any output change of N at the initial real exchange rate, as can be seen in Fig. 5 as well.^[62]

illustration not visible in this excerpt

Fig. 5 (left): Consumption spending effect (a to e) in the market for N

Source: Corden (1984): 360, own thoughts

illustration not visible in this excerpt

Fig. 6 (right): Consumption spending effect (a to e) in the commodity market for N and B & L

Source: Corden / Neary (1982): 829 f., Neary / van Wijnbergen (1986b): 16, Humphrey / Sandbu (2007): 182, own thoughts

Higher domestic real incomes lead to extra expenditure on both traded and non-traded goods, shifting the demand curve for N from D0 to D1 in Fig. 5.^[63] The demand at the initial real exchange rate moves along an income-consumption curve E0s in Fig. 6 until the highest possible indifference curve is reached, and the tangent point d shows the new consumption.^[64]

If some part of the extra income in B is spent, whether directly by factor owners or indirectly through being collected in taxes and then spent by the government, and provided the income elasticity of demand for N is positive, the price of N relative to the prices of the tradeables of L and B must rise (i.e., a real appreciation), as it is set in the domestic market, while the price of traded goods which is determined by international market conditions does not rise despite the extra domestic spending (also referred to as traditional Balassa-Samuelson effect).^[65]

The magnitude of the spending effect is then positively related to the marginal propensity to consume non-tradeable goods, and to the degree that income is spent domestically rather than saved abroad.^[66]

The resulting real appreciation as the endogenous result leads to a structural change, drawing labour out of B and L into N and pushing the output of N.^[67] By the same time, the real appreciation shifts demand away from N towards B and L as a negative substitution effect, since the latter sectors have become relatively cheaper, though the positive income effect due to the boom in B dominates, as point e implies a higher output of N than in point a in Fig. 5 and Fig. 6.^[68]

In Fig. 6, the spending effect’s income effect of demand starts with the change of demand for N at the constant pre-boom relative price of N (from Xn in point bs to Xn in point d). Looking at the income-consumption curve E0s now, it is clear that the actual price of N would have to be much higher in order to stimulate the economy to produce this amount of non-traded goods. For a hypothetical moment, the real exchange rate would be equivalent to the not depicted slope of the tangent budget line in z, the intersection point between E0s and the post-boom production and absorption possibility set. The spending effect’s substitution effect of demand is now the negative consumer reaction to this post-boom hypothetical relative price of N, when the hypothetical price of N is allowed to rise, leading to a budget line rolling westwards along the production possibilities frontier to the new tangent point with the also fallen demand and the absorption possibilities frontier in point e. As a result, the new income-consumption curve E1s can be constructed.

3.1.4 Combining resource movement effect and spending effect

Up to now, the models have been isolated from recursive influences in order to provide a clear picture of the process. But realistically, resource movement and spending effect work together gradually. The commodity market of N in Fig. 7 shows an intuitive view of both resource movement and spending effect by simply adding up graphically the effects explained previously. It is possible to separate both effects as two dashed arrows working at the same time, and still depicting both effects simultaneously in the resulting main arrow from point a to f.^[69]

illustration not visible in this excerpt

Fig. 7 (left): Resource movement & consumption spending effect (a to f) in the market for N

Source: Corden (1984): 360, own thoughts

illustration not visible in this excerpt

Fig. 8 (right): Resource movement & consumption spending effect (a to f) in the commodity market for N and B & L

Source: Corden / Neary (1982): 829 f., Neary / van Wijnbergen (1986b): 16, own thoughts

The picture of Fig. 8 is a synthesis of resource movement effect and spending effect.

As to the effect on output of N at a constant real exchange rate, point brs on the production possibility frontier can be understood as lying between the resource movement effect’s point br north-west to point a in Fig. 4 and the spending effect’s point bs lying north to point a in Fig. 6.^[70]

It would be possible to depict the separate initial income-consumption curves for both effects, but two different equilibriums at a time would distort the logic picture which shall be a clear combination of both effects. Thus, as to the effect on demand for N at a constant exchange rate, the income-consumption curve E0rs represents a mixture of the initial vertical curve E­0r in Fig. 4 and the initial curve E0s in Fig. 6. The resulting intersection point in ‘(ar&d)’ shall only hint at another mixture of the resource movement’s point ar without change in demand and the spending effect’s point d with a positive change in demand, as can be seen in the figures.

The final income consumption curve E1rs includes E1r and E1s­, again in the figures already mentioned.

As to the final equilibrium in f, Xn increases a bit in this paper’s example, as the spending effect is assumed to dominate, but in general, the result is generally ambiguous, as especially Fig. 7 shows intuitively.^[71]

Direct and indirect de-industrialisation . Up to now, the output’s N Sector was in the focus. The following model explains the decrease in output of the L Sector that is presumed to be manufacturing, which thus is called de-industrialisation. It combines both resource movement and spending effect, so some repetitions will be inevitable. The two possible kinds of labour movements out of L, which have been explained in the latter sections, can be categorized as direct and indirect de-industrialisation, depicted in Fig. 9.

illustration not visible in this excerpt

Fig. 9: Direct (L to L’) and indirect (L’ to L’’) de-industrialisation shown in the labour market.

Source: Corden / Neary (1982): 828, 831

The resource movement effect implies the migration out of L and N into B, shifting the labour curve of the Booming Sector leftwards from Lb to L’b in Fig. 9, resulting in a wage rise to w1.^[72] The equilibrium changes from a to br which equals the change of a to br in the previous Fig. 3 and Fig. 4. The resulting decrease of the Lagging Sector’s labour from 0tL to 0tL’ by a direct movement to B is called the direct de-industrialisation.^[73] Mind that though a or br show the same equilibrium in all the figures, the direct de-industrialisation generates only one part of the resource movement effect or the supply shift towards point br in Fig. 3 respectively, as it does not include the migration out of N into B. Nevertheless, de-industrialisation is caused by the resource movement effect alone.^[74]

The more complex indirect de-industrialisation from b to e in Fig. 9 stands for a movement of labour out of L into N and the associated real appreciation resulting both from the reduced output of services (at the initial real exchange rate) due to the resource movement effect, and from the increased demand for services due to the spending effect.^[75] The higher relative price of non-traded goods makes domestic production of traded non-booming goods less attractive, as the behaviour of total factor supply is disadvantageous, and so their output declines.^[76]

For these reasons, the final essential decrease in the labour force from 0tL’ to 0tL’’ with an increase of factor returns to w2 reaching the equilibrium in point f in Fig. 9 must be split.^[77]

The indirect de-industrialisation caused by a part of the resource movement effect – the movement out of N into B as the other part of the above mentioned supply shift towards point br in Fig. 3 – and by the subsequent development from point b to c in Fig. 3 and Fig. 4, reaches an equilibrium in Fig. 9 that is on the Lagging Sector’s labour curve Ll somewhere between b and f. The real appreciation induces the assumed rightward shift of the services sector's labour demand schedule Ln upwards in direction of Ln’ and results in a wage somewhere between w1 and w2.

The indirect de-industrialisation caused by the spending effect does the complementary rest to reach the final equilibrium in point f and can be reconstructed by letting the curves of Lb and Ll coincide in Fig. 9, as if the B Sector did not use any labour, and then by shifting Ln upwards again.^[78]

As a result, the indirect de-industrialisation with a labour movement out of L into N supplements the direct de-industrialisation that resulted from the movement of labour out of L into B.^[79] When de-industrialisation is defined as a fall in output and employment in manufacturing, there must be de-industrialisation in this model provided that there is any spending or resource movement effect.^[80]

It should be stressed that such deindustrialisation, or more general the lagging of sectors, is a symptom of the economy's adjustment to its new equilibrium: it does not provide prima facie grounds for diagnosing a 'disease' that requires corrective action.^[81]

3.2 Differentiating favourable shocks

3.2.1 Non-neutral technological progress

If the technological progress assumed above is not Hicks-neutral anymore, income is still unambiguously raised and thus the spending effect works in the same manner, but the resource movement effect is different.^[82] When capital is assumed to be specific to the energy sector, technological progress can be labour-saving and the sector's demand for labour would decrease at the initial wage.^[83] The resource movement effect can then be reversed, eventually pushing the Booming Sector‘s labour away to the other sectors, and this pushes the output of the Lagging Sector if the technological progress enables the Booming Sector to economise on the factor which it uses intensively relative to the Lagging Sector, resulting in pro-industrialisation in case of manufacturing as the Lagging Sector.^[84] This result is even likely, but the spending effect can be expected to play the dominant role in a typical DD case, so that non-neutral technological progress can still be regarded as a possible cause of the DD.

3.2.2 Resource boom

In many cases the Booming Sector is of an extractive kind, as with minerals, natural gas, crude oil, wood from deforestation, or – apart from raw materials – agricultural land expansion that is often associated with deforestation.^[85] Extractive or not, there is a resource boom which exerts a considerable and mostly sudden windfall. This discovery of valuable resources of any kind corresponds to an increase in the supply of the specific factor within the specific factor model explained above.^[86] Since the oil industry is often almost an enclave, it is realistic to assume that labour markets are barely or not at all affected. The transformation curve in Fig. 4 would shift up less outwards and more strictly upwards, reaching a higher production possibility of traded goods even at the highest N production point.^[87] Then, the resource movement effect would be more difficult to be shown in Fig. 4, and the shift of the supply curve due to labour movement out of N would be smaller in Fig. 3. In the extreme case, if the there is no resource movement effect at all, the transformation curve shifts vertically upwards only.^[88] Thus, for cases of a discovery and depletion of resources where labour is probably barely bid away from other sectors, the resource movement effect is negligible, and the spending effect equals that of the labour-neutral technological improvement shown above.^[89]

This is the main difference from the previously analysed case of a costless technological improvement, but after all, the special issues raised by a natural resource discovery are not necessarily crucial from the point of view of medium-run allocation and distribution problems, since the spending effect brings the clearest symptoms of the DD.^[90]

3.2.3 Price boom of an already produced traded good

A price boom and a resource discovery also do not crucially differ in their consequences, but there are fine differences.^[91] Assume an exogenous rise in the price of a commodity that induces the boom in a sector that exports all of its output:^[92] the main difference is that the production possibilities and thus the amounts of possible products at given relative prices do not rise, but only their values change, differently from the previous scenarios. The resource movement effect is exactly as considered for an improvement in technology, though, since the consequences on profitability and the factor demands of the Booming Sector are the same.^[93] The spending effect, instead, is different, since the real exchange rate now is also changed by the price of B, and a higher price of B generates a substitution effect tending to raise demand for services – assuming plausibly that B and N are net substitutes in consumption, i.e., provided their compensated cross-price elasticity of demand is positive.^[94]

In contrast to a technology improvement as the source of a boom, the increased price of B makes N (and L) relatively cheaper in the isolated view of the substitution effect, decreasing the real exchange rate a bit and increasing the demand for N.^[95]

The isolated income effect starts with increasing demand for N at the pre-boom relative price of N, but in contrast to the technology improvement, the production cannot react immediately and instead the relative price of N skyrockets to the intermediate Pn1 in point g which is vertically above point a in Fig. 10.

illustration not visible in this excerpt

Fig. 10: Consumption spending effect (a to i) in case of a price induced windfall

Source: Gelb (1988): 24, own thoughts

In the previous analysis of a pure spending effect, the substitution effect decreased the demand for N, as was depicted in Fig. 6, and still demand unambiguously grew in the end. A rise in the price as a cause of a boom also comes to this result, and it even accentuates the excess demand for N: both income and substitution effect then lead to an increased demand for N.

This is shown in the following Fig. 10 with the change from point g, the attained point during the income effect when production has not yet reacted, eastwards to point i, working along the new dotted absorption possibility set (not to confuse with the production possibility set) that is gained after the windfall (demarked by the double arrows).

As long as the booming commodity is an export good or the economy is at least a net exporter, the sign of the spending effect is positive in the end, and there is still an unambiguous real appreciation, since the strong appreciation of the income effect dominates the depreciation of the substitution effect.^[96] In Fig. 10, this can be seen by Pn1 as the highest interim real exchange rate (which exists directly after the windfall and before production has reacted) and Pn2 as the final equilibrium real exchange rate, which both must be higher than the initial Pn0, as the steeper slopes illustrate.^[97] The world-wide price boom also affects the trading partners so that the relative price implications of this single-country model are not as straightforward as the implications of a country- or region-specific quantity shock.^[98]

To sum up, this alteration of the previous model is not crucially different, and it gives another tendency to the still ambiguous result of the last sections regarding the change in non-traded goods, as a boom with a windfall component definitely supports the finding that the Non-traded Sector will increase.

3.2.4 Exogenous inflows from abroad

Up to now, the examples assumed that there is a domestic boom and subsequently endogenous inflows. But there is also the possibility of an exogenous increase of inflows from abroad in the form of private and official inflows, be it one-way transfers remittances and plain aid or qualified inflows like official development assistance, increases in the foreign debt, and foreign direct investment.^[99]

One-way transfers have a stronger effect than qualified inflows, as the former correspond to their net present value, while the latter lead to long-run outflows again and thus imply a low net present value which could also be negative, if the net present value of the outflows is higher than that of the inflows.^[100]

For exogenous inflows with a positive net present value which lead to full consumption, there is only a spending effect equalling the spending effect of a price boom and the isolated spending effect of a technological improvement shown above, since the production possibility frontier does not shift, but only the absorption possibility set.^[101] The resource movement effect does not exist, but only an aggregate increase in income by assumption, and the simple reason is that there is no special Booming Sector to which the labour could migrate. There is only the movement from L to N, which was above was the direct consequence of the risen price of N after the resource movement effect, but which happens independently from the resource movement effect in this case.

3.3 Policy of absorption emphasis

The absorption emphasis of the inflows is the most important policy choice. For the intensity of the DD, it is essential whether the spending goes into domestic investment or just into consumption.^[102] Even any form of savings, be it foreign or domestic, will someday result in the absorption decision, unless the possessors decide to spend the money abroad.

It is the choice of governments, companies, and households, but as the government usually controls a considerable part of a favourable shock, be it directly or just indirectly through taxes, it shall be in the focus.

The broad type of absorption can be split in public consumption, public investments, and transfers to the private sector.^[103]

3.3.1 Public consumption

Apart from a typical consumption spending effect as shown for a price boom in Fig. 10, the tendency for more non-traded goods can get even more support when integrating the government as a special consumer. In case of a favourable shock, the inflows often accrue to a great part to the government, and for the example of a price boom, the preliminary model must then be modified by distinguishing the private and the government sector, as in Fig. 11.^[104]

illustration not visible in this excerpt

Fig. 11: Consumption spending effect (a to k) in case of a price induced windfall partly accruing to the government

Source: Gelb (1988): 25

The assumption is that the government spends a fixed proportion of its income on non-traded goods, as marked by the triangles with a one-to-one ratio in this example, in contrast to the private sector which will always adjust its spending on traded and non-traded goods flexibly to the real exchange rate.^[105] The greater the influence of the government then, the less the composite demand for non-traded goods reacts to a change in the real exchange rate. In the extreme case of a government that absorbs the full windfall gains, the substitution effect does not change the real exchange rate anymore, so that all additional spending of the boom windfalls is adjusted only by the income effect. Then, the real exchange rate is not the result of the flexible market anymore, but the result of the governmental fixed proportion in spending.

The higher the fixed proportion of spending on N, the more appreciated is the exchange rate, as the example with point k lying east to point i in spite of the same Pn1 adumbrates, and the other way round, if the fixed proportion were more in favour of the traded goods, there could be less real appreciation than in point i in spite of the same Pn1. In addition, the greater government’s fixed proportion of demand for N, the more real appreciation takes place in the early stages of the boom, before the economy has been able to reallocate resources to the Non-traded Sectors.^[106] Normally, in later stages of for example a resource boom, the real exchange rate depreciates from its peak when resources get reallocated so that the cost of purchases by the public sector relative to the booming commodity prices gets reduced.^[107] But in the extreme case when the government spends the full windfall only on non-traded goods as a fixed proportion, the initial real exchange rate Pn0 is immediately appreciated to a vertical line as the highest possible, and it will not depreciate since the production maximum of Xn is reached only at the maximum appreciation. As a consequence, the domestically produced traded goods are crowded out completely (point l), while the excessive rest of the windfall gains and the money that the producer earn will be spent on imported goods (point m).

Of course, these extreme cases are unrealistic, but one can conclude at least that the extent of a real appreciation is greater, the more intensive the nontraded good is in public demand.^[108] This is actually a conclusion that can be applied to the spending of the private sector, too, and it was mentioned before by integrating the marginal propensity to consume N. But the difference is that the private sector usually has a more flexible proportion of N in its consumption basket, and more, that the change of private spending can be expected to be less biased towards N than the change of governmental spending, while the government can be expected to spend a greater proportion on N especially at higher windfalls, as it can exert a much wider range of services and the demand for traded goods will be satisfied more quickly. Typical consumptive government activities provide state administration and national and public security, residential and municipal services or general public service, all with respect to their consumptive part.^[109]

3.3.2 Public domestic investment

The analysis of public domestic investment can be generalised to investment as such, as the private sector’s absorption will be split into consumption and domestic investment.

For the emphasis of the non-booming Traded Sectors, the government can take part in the private sector, be it as a possessor of companies or as an investor supporting the private sector, for example in sectors like food and nonfood agriculture, natural resource processing, and manufacturing.^[110]

In case of a pure investment into a non-booming Traded Sector, the transformation curve shifts upwards only with the higher productivity reaching a higher output of traded goods, as analysed before in the context of the technological improvement.^[111]

The marginal product of labour in the Booming Sector then rises, subsequently leading to higher real wages, higher income, higher aggregate demand which aims in part on non-traded goods, real appreciation, and the contraction of the non-booming Traded Sector (which is enforced by the parallel labour movement into the Booming Sector, if possible).^[112] Hence, investment does not seem to change the model’s basic argumentation chain explained before.^[113]

But it is not likely that the non-booming Traded Sector is the investment target alone, since the Non-traded Sector is often closely connected with the Booming Sector and, what is more, the majority of governmental investment goods is non-traded, facilitating the traded sectors to produce their products more efficiently.

Most governments emphasize key infrastructural sectors as productivity enhancing public goods in their investment program, goods that are generally non-traded.^[114] Such goods can be physical capital largely serving production (transport, communications, and power supply), physical "social" capital (housing) and human capital (education, health).^[115] They can be regarded as investments as long as they are combined with an expected productivity enhancement or other stimulation effect in the future, but they do not themselves create activities that can substitute for the foreign exchange provided by the Booming Sector.^[116] The consecutive productivity enhancement counteracts the spending effect on domestic goods, and this is why a good investment program is so efficient against DD symptoms.^[117]

An investment can have ambiguous effects both on output and prices, so there is a good side and a dark side again.

Invested windfalls offsetting or mitigating the DD. Investments usually have a positive impact on productivity, which would change the preliminary result of the assumed pure consumption scenario that the production possibility frontier does not move after a price induced boom. Indeed, forestalling the main conclusion, a proper attitude towards investments can mitigate or even prevent the symptoms of the DD.^[118]

Mind that the transformation curve NT of the following models now depicts the situation when the windfall has already been earned and waits to be spent. It could be spent on consumption, with the consequent absorption possibility set shown in Fig. 10, but it could also be fully invested with the consequently higher production possibility frontier in the next Fig. 12 which then equals the higher absorption possibility frontier.^[119]

For a simple beginning, the implied assumption is that the windfall is spent fully on imported investment goods in Fig. 12, hence excluding a direct spending effect.^[120]

illustration not visible in this excerpt

Fig. 12: Production and absorption possibilities change when the windfall is fully spent on imported investment goods

Source: Sachs (2007): 185

The production and absorption possibilities frontier then shifts outward (i.e. upward and rightward rather than simply rightward), and in the example, production and consumption of both non-booming traded goods and non-traded goods increase equally.^[121] The actual rise of both frontiers depends on the efficiency of the investments, but final consumption must share its place with the additional investment in this rather short-run model.^[122]

As the direct spending effect does not work in this example, the demand is no longer an important determinant of the first-stage change of the real exchange rate, and consequently the real exchange rate itself has lost its impact on output, and the output is now determined by the change in productivity.^[123] There is an indirect spending effect from the higher productivity in the Non-traded Sector, though, and this both tends to increase the wages (thus the relative prices of non-traded goods), and by the same time this price increase of the non-traded goods can be dominated by the price reduction due to the risen output. A real depreciation is possible, but it does not matter that much anymore.^[124]

In this example with a straight income-consumption curve Ei and a parallel outward shift of the frontiers the real exchange rate does not change at all. But it is perfectly possible that the exchange rate rises, and still the non-booming tradeable production expands, which would correspond to a bent income-consumption curve or to a rightwards shift of the income-consumption curve.^[125]

Now the scenario of a windfall partly spent on non-traded investment goods is analysed, with the difference of an additional spending effect. At a given resource discovery that has to be depleted, inflows will often induce investments biased towards the Booming Sector and its non-traded linked sectors. But there are distorting side-effects of investment, as the spending on non-traded goods with its subsequent real appreciation can be fully outbalanced by the improved Non-traded Sector’s productivity, higher output and lower price, implying real depreciation. The consequences on the real exchange rate can differ from case to case.

The extreme case of a windfall fully spent on non-traded investment goods can best explain the ambiguity of the process. As investments need time to spread their advantage, a short-term real appreciation and the associated squeeze of the traded sector can happen in the first stages of a boom as long as the productivity enhancing effects both on non-tradeable and tradeable production are still to come.^[126] This transitory phase is likely to be very short-lived when pushing demand for non-traded goods artificially high before the productivity effects have come to work, as shown in Fig. 13 and Fig. 14 by the dotted budget line P1.^[127]

The spending effect is mitigated by the next step that includes the higher productivity of the Non-traded Sector and shifts the transformation curve to N’T in both figures. This stage is still isolated from a possible productivity improvement of the traded sector.^[128] The result can even culminate in a decrease in welfare, which is not shown in the pictures. At this second stage, the initial real appreciation from P0 to P1 is reversed again.

illustration not visible in this excerpt

Fig. 13 (left): Full investment into the Non-traded Sector without productivity enhancement of the traded sector

Source: Sachs (2007): 186

illustration not visible in this excerpt

Fig. 14 (right): Full investment into the Non-traded Sector with non-traded goods turning into tradeables and productivity enhancement of the traded sector

Source: Sachs (2007): 186, own thoughts

Poor countries with a lack of essential non-traded services can likely incur a real depreciation if the public investment financed by the windfall substantially raises the productivity of the Non-traded Sector.^[129] An example could be the case of poor farmers who produce non-tradeable food and take advantage of a governmental investment into irrigation, so that the increase in their non-traded output reduces the price of non-tradeables, implying a real depreciation.^[130]

Even in a scenario of a high positive external effect on the productivity of the traded sector a real depreciation is still possible, as shown in Fig. 14 by the final transformation curve N’T’ and the associated final real depreciation in P3. It could be an example of a smallholder farmer whose productivity increases due to new infrastructure and thus gets first access to the markets so that his non-traded food becomes traded. When the rest of produced non-tradeables additionally increases in output, it is possible that that there is real depreciation and a rise in the output of tradeables.^[131] It is thus likely that both the short-term appreciation and the long-term depreciation occur in a poor country, and the real exchange rate might not change after all. The second-stage depreciation process mitigates the DD, and it is clear that intelligent investment changing the non-traded goods into traded goods could be the solution to prevent symptoms of the DD, as such change can dominate the spending effect.

One might speculate at this point that in contrast to the poor countries, rich countries are more likely to incur a real appreciation which decreases the traded sector, as the productivity of the Non-traded Sector usually cannot be so strongly enhanced and the spending effect consequently has more power.

Invested windfalls exacerbating the DD. Coming to the dark side, a bad investment program into the Non-traded Sector shall explain how the DD could even be intensified due to investments, which has happened to many developing countries that have exerted very ambitious but unsuccessful governmental investment programs.

“If the increased spending (...) falls partly on non-traded goods (rather than entirely on imported capital goods), there can be a minor DD effect emanating (…) from the investment boom itself.”^[132] And there is a scenario within the previous specific factor model that very likely supports real appreciation with a disadvantage for the Lagging Sector at a given investment into the Non-traded Sector, and that is overinvestment into non-traded investment goods.^[133] The contraction of the traded sector is most likely if this investment increases the productivity of the Non-traded Sector more than the productivity of the traded sector which will usually profit from investments into the Non-traded Sector, too (to be seen in Fig. 15 by the outward, but more rightward shift of the transformation curve), and the simultaneous second requirement is that the investment in the Non-traded Sector is so strong that a part cannot be turned into productivity growth of the Non-traded Sector but instead stimulates the non-traded goods’ prices again, as often happens in the real estate sector or infrastructural projects (see the higher slope in P2 than in P0, equalling a real appreciation in the final equilibrium).^[134]

illustration not visible in this excerpt

Fig. 15: Investment leading to real appreciation and a Lagging Sector

Source: Gelb (1988): 24, own thoughts

The restrictions are necessary, since the investment in the Non-traded Sector can as well spill over strong productivity enhancements to the traded sectors which in poor countries can plausibly reverse the unfavourable sectoral shifts and induce pro-industrialisation (with the real exchange rate independent of this particular shift), when shifting the transformation curve N’T’ more and more upwards in Fig. 15, as was shown in Fig. 14.

To recapitulate, the impact of investments into the Non-traded Sector on the real exchange rate and on sectoral shifts is rather ambiguous, and still, it is not implausible to assume that very strong investments of any cause can lead to real appreciation and deindustrialisation, if non-traded investment goods are too strongly demanded, so that the price increase due to a higher demand likely dominates the price decrease due to higher output of non-traded goods. Especially for a price boom this outcome is likely, since the government will try to raise investment projects in the Non-traded Sector which will often end up in overinvestment due to unnecessary, unnecessarily expensive and huge projects as it is often perceived in developing economies. The government need not be the main cause, as exaggerated and inefficient foreign direct investments into the Non-traded Sector can also plausibly happen during the development of first oil fields or in order to gain first essential market shares.

3.3.3 Transfers to the private sector

The absorption through transfers to the private sector usually entails both investment and consumption, for example by reducing non-oil taxes, granting subsidies to selected activities or groups – in particular the Lagging Sector – or subsidizing certain products.^[135]

3.4 Extensions of the Core Model

3.4.1 Labour market disequilibrium regimes

The response to the surge in demand will create disequilibrium in some markets at pre-windfall output and prices, and an important disequilibrium is in the labour market.^[136] There are two cases which can lead to unemployment: real wage rigidity and nominal rigidity.^[137] Up to now, the assumption that all factor prices and the price of N are flexible has ruled out unemployment.^[138] But in reality, the economy does not necessarily move smoothly and instantaneously to the new equilibrium and short run agents on the long side of either market can be rationed.^[139]

Different exogenous values of the wage and the price of N give rise to different disequilibrium regimes, which can be depicted in a space of wages related to real appreciation as in Fig. 16.^[140]

illustration not visible in this excerpt

Fig. 16 (left): Disequilibrium regimes with wage and price rigidities

illustration not visible in this excerpt

Source: Neary / van Wijnbergen (1986b): 20

Fig. 17 (right): Effects of a boom with wage and price rigidities

Source: Neary / van Wijnbergen (1986b): 22

This is the basic model of the disequilibrium regimes generated from wage and price rigidities, and the regimes are labelled C for classical unemployment (too small capital stock, give the chosen degree of capital intensity, to employ all workers, excess demand for consumer goods due to high money balances, too high nominal wages = too low labour demand, too high real wages = excess supply of labour), K for Keynesian unemployment (too high nominal wages = too low labour demand, too low real wages = too low consumption) and R for repressed inflation (excess demand for goods, high capital per worker, high level of real money balances per person, too low nominal wages = excess demand for labour, too low real wages = too low labour supply, ‘labour shortage’ ).^[141]

The borders between the regimes can be constructed as follows. The westward line from point a shows the labour market equilibrium locus when the labour supply is the exogenous determinant of employment.^[142] The employment locus eastwards from point a must be downward-sloping, as the lower demand for N deriving from the higher relative price of N (i.e. a higher q) leads to excess supply of N, and the therefore rationed producers scale down their labour demand and pay lower wages, the more southwards the locus goes.^[143] Finally, the line upwards from point a depicts the labour market equilibrium locus when income is no longer exogenously determined by full employment, as is with the westward line, but endogenously determined by the combination of both output and an already included value of the natural resource discovery as the representative example (the latter making the line a bit steeper).^[144]

Classical unemployment and repressed inflation can now occur when the boom shifts the loci to the right as in Fig. 17, while the Walrasian equilibrium in point d is only reached after an adjustment phase, exerting a short-run disequilibrium in point a.^[145]

As the repressed inflation is no unemployment, but labour shortage, it is explained separately afterwards.

The Keynesian unemployment, though, does not occur in this context but arises from the monetary policy of money supply rigidities. It thus will be analysed within the monetary type of the DD.

3.4.1.1 Classical unemployment

There are two conditions of how classical unemployment can occur. There must be some degree of wage rigidity, and additionally, the weight of non-traded goods in the consumption basket of wage-earners must be sufficiently large.^[146]

At a given resource discovery, the higher demand then leads to an expansion of the nontraded sectors (the first stage, before supply-side effects of possible investment increase capacity, is meant here), requiring a rise in the relative price of the non-tradeds and the consecutive contraction of the non-booming Traded Sectors.^[147] The question is how quickly the total supply reacts to the increased demand, so that the inflexibility of labour is the determinant of the adjustment problem, and the more inelastic the labour is the more additional real appreciation will be generated at a given wage.^[148]

This adjustment process does not necessarily lead to unemployment. The actual reason for transitory unemployment is the standard augmented Phillips curve assumption that real consumption wages can only be reduced by temporary unemployment.^[149] The wage indexation locus WI in Fig. 17 depicts the real consumption wage that is identified with the utility level of wage-earners, and by invoking their expenditure function.^[150] The larger the share of the non-traded good in the consumption basket of wage-earners, the greater is the slope of the WI locus.^[151] In the example, the slope of WI is steeper than the line from point a to the new equilibrium in point d, which simply implies that on the way to point d, the real consumption wage becomes continuously too high (shown by the three arrows directed southwards), so that it has to fall and transitional classical unemployment must result according to the augmented Phillips curve assumption mentioned above.^[152]

For simplicity, this analysis has assumed a resource boom that does not use any labour, which often mirrors the reality quite well, so that the spending effect as the main effect regarding the DD is active alone; with a resource movement effect, the necessary contraction of the non-booming Traded Sectors will be stronger the more labour has to be given up to the Booming Sectors.^[153]

Up to now, the analysis has dealt with the total labour supply that is intersectorally mobile. But there is an important case where the labour supply is partly sector-specific and generates sector-specific unemployment. Such unemployment can arise when some types of labour are specific to the Lagging Sector in the medium-run at least, and practise real wage resistance.^[154] Even with no inter-sectoral movements of the specific factors, the real wage would have to fall both because of the movement of mobile factors into other sectors and because of the rise in the price of N, so that the DD then manifests itself, partly at least, in extra unemployment of such factors.^[155] The unemployment would even be intensified if workers in the Lagging Sector adjusted their real wage demand to the real wages of workers specific to the Booming Sector, where market forces will have raised real wages.^[156]

3.4.1.2 Repressed inflation – labour shortage

For the case of repressed inflation, the WI locus is, in contrast, less steeply sloped than the line from point a to d in Fig. 17, so that the real consumption wage becomes too low on the way to point d (transitory labour shortage) and must rise.^[157] In this scenario, the slope of the WI locus implies a relatively too low demand for N to meet the equilibrium, or in other words, the non-traded good sector is less important in demand than in supply.^[158] The solution is to have both an increase in the wages and inflation (which only affects the non-tradeds by assumption) – that is where the name ‘repressed inflation’ comes from.

Labour shortage due to a real consumption wage that is lower than the real wage can for example occur when the preliminary assumption of no consumption-leisure trade-off in the household utility function is removed and the income effect dominates the substitution effect of labour supply.^[159] This scenario happens in particular in very poor countries or when favourable shocks finance high unemployment and other benefit levels, in turn raising voluntary unemployment.^[160] Labour is then partly substituted with leisure, and the higher reservation wage is equivalent to a lower real consumption wage (that is a lower utility at the given wage rate) – the result is labour shortage.

The relatively lower slope of WI implies that the consumers tend to reduce the share of N in their consumption basket.^[161] This bias is realistic, since a household normally has limited utility from spending its additional income on non-traded goods, while the traded goods can raise the utility relatively better.

The existence of a resource movement effect makes it more likely that labour shortage rather than classical unemployment will emerge, since it raises the wage rate in addition to the rise from the spending effect, so that the possible distance between equilibrating wage and real consumption wage in Fig. 17 grows.^[162]

3.4.2 Price boom and domestic use of the booming good

So far, the booming good has been considered as a source of foreign exchange only, and not as a domestically absorbed product.^[163]

When the assumption is given up that the Booming Sector product is wholly exported, the Core Model faces a domestic absorption effect – namely the effect of the price rise of a product that is produced and sold at home, even though the actual price may be determined in the world market.^[164]

The domestic absorption effect can be split into the extreme cases of using the booming product as a full domestic consumption good or as an intermediate good.

Forestalling the result, the spending effect will be ambiguous in both cases and thus weakens the main DD process, and though the stronger resource movement effect decreases the Lagging Sector, as it is typical of the DD, there is no paradoxical nature in it, as a domestic price rise is simply an unfavourable shock when the products are fully absorbed in the home market.

It is useful to keep the domestic absorption effect separate from the main analysis, since in reality it is not usual that there is a full domestic absorption of the booming good though it is actually a very normal case that a country does not only export, but also use its products. Resource-rich countries are normally rather net exporters, and the DD thus does not lose its relevance due to these distortions. In addition, these effects often lose their intensity when the policy option is chosen to subsidize the growing input prices in order to support the non-booming sectors.^[165]

Booming Sector product is final consumption good. When the booming product is a final consumption good, the spending effect gets both positive out of the higher income of the Booming Sector and negative out of the lower income of the other sectors.^[166] Only if the marginal propensities to spend on N differed, the net spending effect as a result of the domestic absorption effect is not neutral.^[167] Coming from the spending effect’s income effect to the substitution effect, the higher price of B faced by domestic consumers may lead to substitution in favour of L and N, this increasing the likelihood of appreciation.^[168]

The resource movement effect is the same as in the Core Model: the marginal product of labour in B will initially rise relative to that in L and N, but the resource movement effect on the home price rise must be added to the resource movement effect of the export price rise, so that the basic resource movement effect is enforced.^[169] It is possible, though, that the real appreciation is reduced again when the factor-intensities in the sectors differ between L and N. For the case of oil, output of N at constant prices of N will tend to rise since N is normally relatively less oil-intensive, and in that case the real appreciation will finally be less.^[170]

Price boom of an intermediate input. If the Booming Sector product is an input in L and N rather than a final consumption good, consumers are not directly affected, even though they will be indirectly affected through a change in the price of N.^[171] There is a positive spending effect out of the Booming Sector, and simultaneously a negative one from the fall in other sectors’ factors, since the higher price of oil squeezes the values added per unit (the effective prices, as distinct from the nominal prices) in L and N at a constant price of N.^[172] The substitution effect of the spending effect will tend to substitute B for L and N, this adding to the rise in demand for N and thus the tendency to real appreciation.^[173]

[...]

^[1] cf. Corden (1984): 359; cf. Gelb (1988): 22; cf. Lartey (2007): 339; the term was coined by the Economist on November 26, 1977, p. 82 f.

^[2] cf. Frijns (1986): 136 f.; cf. Corden (1984): 359 and Fn. 1.

^[3] cf. Neary / van Wijnbergen (1986a): 3.

^[4] cf. Neary / van Wijnbergen (1986b): 14.

^[5] cf. Sachs (2007): 175; cf. Usui (1997): 151.

^[6] cf. Sachs (2007): 175; cf. López / Molina / Bussolo (2008): 218; cf. Usui (1997): 151; the higher governmental financial power is essentially needed in structurally instable regions.

^[7] cf. Sachs (2007): 177.

^[8] cf. Usui (1997): 151.

^[9] cf. Usui (1997): 151.

^[10] cf. Corden / Neary (1982): 825 Fn. 1.

^[11] As the optimal treatment of the disease strongly varies between the individual cases, the more neutral French term 'Syndrome Hollandais' might be more adequate, cf. Neary / van Wijnbergen (1986b): 14 and Fn. 3.

^[12] cf. Neary / van Wijnbergen (1986b): 13.

^[13] cf. Gelb (1988): 10; probably first Booming Sector analysis in 1857 by John Cairnes, cf. Gelb (1988): 21; The other way round, it can as well be a favourable shock that leads to a Booming Sector due to a reallocation of productive factors, cf. Gelb (1988): 22.

^[14] cf. Neary (1986): 324 f.

^[15] cf. Neary (1986): 325.

^[16] cf. Neary (1986): 325.

^[17] cf. Neary (1986): 325.

^[18] cf. Neary (1986): 326; DD consequences on the Spanish industry of the inflow of American treasure in the sixteenth century, gold discoveries in Australia in 1851 with subsequent DD, cf. Corden (1984): 359; cf. Gelb (1988): 21.

^[19] cf. Corden (1984): 360.

^[20] cf. Corden / Neary (1982): 825.

^[21] cf. Corden (1984): 362; cf. Fetisov (2007): 54.

^[22] cf. Neary (1986): 331.

^[23] cf. Neary / van Wijnbergen (1986b): 13.

^[24] cf. Neary / van Wijnbergen (1986b): 13.

^[25] cf. Neary (1986): 331.

^[26] DD is an intrinsically dynamic story according to Kang / Prati / Rebucci (2007).

^[27] cf. Corden / Neary (1982): 825.

^[28] cf. Corden / Neary (1982): 827; cf. Neary / van Wijnbergen (1986a): 2.

^[29] cf. Gelb (1988): 22.

^[30] cf. Neary / van Wijnbergen (1986b): 17; cf. Corden (1984): 362; an example of a dominating important resource movement effect is Australia’s mineral boom in the oil price boom decade after 1973, cf. Forsyth (1986), but this is an exception.

^[31] cf. Corden (1984): 377; cf. Corden / Neary (1982): 827; some examples of an enclave: Britain, cf. Corden / Neary (1982): 827; Indonesia and Mexico, cf. Usui (1997): 151 Fn. 1, e.g. employment of the Indonesian oil sector was less than 1 per cent of the total workforce so that spending effects dominated resource movement effects, cf. Neary / van Wijnbergen (1986a): 8.

^[32] cf. Neary / van Wijnbergen (1986b): 14.

^[33] cf. Neary / van Wijnbergen (1986b): 34.

^[34] cf. Neary / van Wijnbergen (1986b): 13 f.

^[35] cf. Corden (1984): 371.

^[36] cf. Corden (1984): 371.

^[37] cf. Corden / Neary (1982): 825 f.

^[38] cf. Neary / van Wijnbergen (1986b): 13.

^[39] cf. Corden / Neary (1982): 833; cf. Neary / van Wijnbergen (1986b): 17.

^[40] cf. Corden / Neary (1982): 826, and Fn. 1 for discovery and technological improvement; cf. Neary / van Wijnbergen (1986b): 14 and cf. Corden / Neary (1982): 840 for discovery and price; cf. Corden (1984): 360 for price and transfers (only spending effect).

^[41] cf. Corden / Neary (1982): 825 offer the respective country examples; cf. Barbier (2004): 137.

^[42] cf. Neary / van Wijnbergen (1986b): 17; following the Ricardo-Viner model (one mobile factor combined with a specific factor in each industry) that was revived by Jones (1971).

^[43] cf. Corden / Neary (1982): 827.

^[44] cf. Neary / van Wijnbergen (1986b): 14; example: B = energy, L = manufacturing, N = Services cf. Corden / Neary (1982): 826.

^[45] cf. Corden / Neary (1982): 826.

^[46] cf. López / Molina / Bussolo (2008): 218.

^[47] cf. Corden / Neary (1982): 827.

^[48] This assumption rules out the possibilitity of 'immiserising growth' for the economy as a whole cf. Corden / Neary (1982): 826.

^[49] cf. López / Molina / Bussolo (2008): 218; Workers do not exchange working time for leisure when they earn more money.

^[50] cf. Corden / Neary (1982): 826.

^[51] cf. Corden (1984): 360 and own thoughts on the basis of cf. Corden (1984): 360 Fn. 4; The terms of trade are assumed fixed, both tradeables (B and L) can be aggregated into a single Hicksian composite traded good, Xt cf. Corden / Neary (1982): 829.

^[52] cf. Corden (1984): 360 and Fn. 4.

^[53] cf. Corden (1984): 360 f.; cf. López / Molina / Bussolo (2008): 218; the demand shifts by an amount proportional to the extent of a technological progress or a price increase cf. Corden / Neary (1982): 830; Supply and thus the output of each sector depends on the real product wage it faces, cf. Neary / van Wijnbergen (1986b): 17.

^[54] cf. Corden (1984): 361, cf. Corden / Neary (1982): 830.

^[55] cf. Corden / Neary (1982): 830.

^[56] cf. Corden / Neary (1982): 829 f.

^[57] cf. Corden / Neary (1982): 830.

^[58] cf. Corden / Neary (1982): 830; cf. Corden (1984): 361.

^[59] PB is assumed to be fixed for this case of a technological improvement.

^[60] The fall of Xn cannot be reversed into a rise, though cf. Corden / Neary (1982): 830; cf. Corden (1984): 361; cf. López / Molina / Bussolo (2008): 218; cf. Neary / van Wijnbergen (1986a): 2.

^[61] cf. Sachs (2007): 184 f.

^[62] cf. Corden / Neary (1982): 830; Point bs stands for the ‘b’ of the spending effect in contrast to br, the ‘b’ of the resource movement effect that lies northwest of point a. Due to this distinction all the points are still comparable among the figures.

^[63] cf. Corden / Neary (1982): 830; cf. Neary / van Wijnbergen (1986a): 2; cf. Fetisov (2007): 54.

^[64] cf. Corden / Neary (1982): 830.

^[65] cf. Corden (1984): 360; cf. Corden / Neary (1982): 827, 830; cf. Neary / van Wijnbergen (1986a): 2; Sachs (2007): 182; For example, most Latin American countries are price takers in international markets, growing demand does not raise the prices of tradeables, cf. López / Molina / Bussolo (2008): 218; Non-tradeables cannot be imported and thus do not have to face competition and the market clears by domestic price movements, while tadeables obey the law of one price, cf. Gelb (1988): 22; cf. López / Molina / Bussolo (2008): 223.

^[66] cf. Corden / Neary (1982): 828; cf. Gelb (1988): 22.

^[67] cf. Corden (1984): 360; cf. Fetisov (2007): 54; cf. Usui (1997): 151; cf. Corden / Neary (1982): 830 f.; This labour movement due to a real appreciation will be referred to as indirect de-industrialisation in the next section.

^[68] cf. Corden (1984): 360; cf. Corden / Neary (1982): 831.

^[69] The actual direction of the effects would be along the supply curves from a to c and e, respectively, showing a kind of parallelogram of forces.

^[70] As the transformation curve has not shifted vertically upwards, the expression is not exact.

^[71] cf. Corden / Neary (1982): 831.

^[72] With the increase of labour demand in the Booming Sector, the composite labour demand schedule shifts upwards, cf. Corden / Neary (1982): 830.

^[73] This movement does not involve the market for N and thus does not require an appreciation of the real exchange rate, cf. Corden (1984): 361.

^[74] cf. Corden / Neary (1982): 831.

^[75] cf. Corden (1984): 361; cf. Corden / Neary (1982): 831.

^[76] cf. Neary / van Wijnbergen (1986a): 2; cf. López / Molina / Bussolo (2008): 218.

^[77] cf. Corden / Neary (1982): 831; cf. López / Molina / Bussolo (2008): 218.

^[78] cf. Corden / Neary (1982): 830 f.

^[79] cf. Corden (1984): 361; cf. Fetisov (2007): 54.

^[80] cf. Corden / Neary (1982): 832.

^[81] cf. Neary / van Wijnbergen (1986b): 41.

^[82] cf. Corden / Neary (1982): 839.

^[83] cf. Corden / Neary (1982): 839.

^[84] cf. Corden / Neary (1982): 840.

^[85] cf. Corden / Neary (1982): 825 offers the respective country examples; cf. Barbier (2004): 137.

^[86] cf. Corden (1984): 360; the additional production can be achieved without additional investment.

^[87] cf. Sachs (2007): 182.

^[88] cf. Corden / Neary (1982): 830.

^[89] Output of N must then finally be higher than in the pre-boom situation, provided spending on non-tradeables goes up initially. cf. Corden (1984): 362.

^[90] cf. Corden / Neary (1982): 826 Fn. 1.

^[91] cf. Corden / Neary (1982): 840; cf. Neary / van Wijnbergen (1986b): 13.

^[92] cf. Corden (1984): 360.

^[93] cf. Corden / Neary (1982): 840.

^[94] Generally, a price induced boom affects national income differently from a technology improvement, too, cf. Corden / Neary (1982): 840.

^[95] cf. Corden / Neary (1982): 840.

^[96] cf. Corden / Neary (1982): 840.

^[97] Note that Pn1 going through point h shall not be understood as the result of a production change, as the preliminary real appreciation to Pn1 only induces the production to shift to point h cf. Gelb (1988): 24.

^[98] cf. Neary (1986): 331; cf. Neary / van Wijnbergen (1986a): 2.

^[99] cf. Lartey (2007): 338.

^[100] This is generally the case in a perfect capital market and also counts for the particular case of debts if the interest on the debt equals the attained interest from the borrowed money.

^[101] cf. Neary / van Wijnbergen (1986b): 14; strong similarity to a resource boom, cf. López / Molina / Bussolo (2008): 218.

^[102] cf. Corden (1984): 372; it can also go into foreign investment, but this only postpones the decision how to absorb the money domestically.

^[103] cf. Gelb (1988): 40.

^[104] cf. Gelb (1988): 25 f.

^[105] cf. Gelb (1988): 26; It is of course unrealistic to either assume for the private sector a complete flexibility of spending or a fixed proportion spent on non-traded goods, since at least a basic amount of non-traded goods is certainly needed and the proportion spent on N can be expected to change only above that level.

^[106] cf. Gelb (1988): 26.

^[107] cf. Gelb (1988): 26.

^[108] cf. Gelb (1988): 26.

^[109] cf. Fetisov (2007): 54 ff.

^[110] cf. Gelb (1988): 40.

^[111] A rise in the productivity in the Booming Sector would construct the equivalent picture. cf. Corden / Neary (1982): 827.

^[112] cf. Lartey (2007): 339.

^[113] There are many sub-options like large-scale and capital-intensive or small-scale and labour-intensive activities, but the scenario is held as simple as possible. cf. Gelb (1988): 40.

^[114] cf. Gelb (1988): 41.

^[115] cf. Gelb (1988): 41.

^[116] cf. Gelb (1988): 41.

^[117] The time profile of non-booming fiscal revenues and expenditures is also affected, but this is a dynamic issue, cf. Gelb (1988): 41.

^[118] cf. López / Molina / Bussolo (2008): 223.

^[119] cf. Sachs (2007): 184.

^[120] cf. Sachs (2007): 184.

^[121] cf. Sachs (2007): 184, for example due to new infrastructure or better education.

^[122] The exact shift of the absorption possibilities frontier depends on the quality of the investment according to its net present value of the expected returns and the functioning of the capital markets.

^[123] cf. Sachs (2007): 184.

^[124] cf. Sachs (2007): 184.

^[125] cf. Sachs (2007): 187.

^[126] cf. Sachs (2007): 187.

^[127] cf. Sachs (2007): 184; the companies anticipate that investments need some time and accept relative losses for this transitory phase.

^[128] cf. Sachs (2007): 184.

^[129] cf. Sachs (2007): 185.

^[130] own thoughts on the basis of cf. Sachs (2007): 185.

^[131] own thoughts on the basis of cf. Sachs (2007): 185.

^[132] cf. Sachs (2007): 184.

^[133] cf. López / Molina / Bussolo (2008): 219 f.; the rather DD-neutral case of imported investment goods is explained in section 3.4.

^[134] Prices of financial assets, and particularly of real estate, can rise rapidly and affect the sectoral allocation of investment, cf. López / Molina / Bussolo (2008): 219 f.; these processes also attract speculative investments.

^[135] cf. Gelb (1988): 40.

^[136] cf. Gelb (1988): 37.

^[137] cf. Neary / van Wijnbergen (1986a): 2.

^[138] cf. Corden (1984): 369.

^[139] cf. Neary / van Wijnbergen (1986b): 19.

^[140] cf. Neary / van Wijnbergen (1986b): 19.

^[141] cf. Neary / van Wijnbergen (1986b): 21; cf. Gelb (1988): 163 f.

^[142] This assumes that no agents are ever rationed in the market for traded goods, and since labour supply is exogenous, the locus is unaffected if households are rationed in the non-traded goods market, cf. Neary / van Wijnbergen (1986b): 19 f.

^[143] Under full employment, the effective non-traded good market equilibrium locus goes the same way, cf. Neary / van Wijnbergen (1986b): 20.

^[144] cf. Neary / van Wijnbergen (1986b): 21.

^[145] cf. Neary / van Wijnbergen (1986b): 21.

^[146] cf. Neary / van Wijnbergen (1986b): 41.

^[147] cf. Gelb (1988): 42.

^[148] cf. Gelb (1988): 42.

^[149] cf. Neary / van Wijnbergen (1986b): 21 f.

^[150] This is only the standard wage indexation rule about the behaviour of the wage rate cf. Neary / van Wijnbergen (1986b): 22.

^[151] cf. Neary / van Wijnbergen (1986b): 22.

^[152] cf. Neary / van Wijnbergen (1986b): 22 f.

^[153] cf. Neary / van Wijnbergen (1986b): 19; cf. Gelb (1988): 42.

^[154] cf. Corden (1984): 370.

^[155] cf. Corden (1984): 370.

^[156] cf. Corden (1984): 370.

^[157] This follows indirectly from Neary / van Wijnbergen (1986b): 22 f.; Points below and to the right of point a are irrelevant to the effects of natural resource exploitation, the slope of WI cannot be negative if N is a normal good. This follows indirectly from Neary / van Wijnbergen (1986b): 21 Fn. 9.

^[158] cf. Neary / van Wijnbergen (1986b): 23.

^[159] cf. López / Molina / Bussolo (2008): 218.

^[160] see the Netherlands as example, so the DD is Dutch in this point, cf. Corden (1984): 369 Fn. 14.

^[161] cf. Neary / van Wijnbergen (1986b): 23.

^[162] cf. Neary / van Wijnbergen (1986b): 23.

^[163] cf. Gelb (1988): 27.

^[164] cf. Corden (1984): 368.

^[165] cf. Corden (1984): 368; this policy reduces the tendency for the exporters' real effective exchange rates (explained in 7.1.2) to appreciate against those of their trading partners, cf. Gelb (1988): 27.

^[166] cf. Corden (1984): 368.

^[167] For the following analysis, the assumption is still that domestic prices always move with world prices, cf. Corden (1984): 368.

^[168] cf. Corden (1984): 368.

^[169] cf. Corden (1984): 369.

^[170] cf. Corden (1984): 369.

^[171] cf. Corden (1984): 369.

^[172] cf. Corden / Neary (1982): 840; cf. Corden (1984): 369.

^[173] cf. Corden (1984): 369.