This article aims to critically analyze the Scandinavian model of inflation, considering the recent contributions of Sraffian approach to inflation theory. Besides a short introduction describing relevant historical conditions, the article is organized in four other sections followed by closing remarks. Section 2 presents the Scandinavian model in its original version and policy implications pointed by its authors. The following section discusses the limitations of the Scandinavian model analysis of distribution. By its turn, section 4 and 5 present alternative approaches to solve the model, flexing some of its hypothesis and obtaining new conclusions. Finally, we briefly discuss the main results in the conclusion section.

During the Golden Age, the advanced capitalist economies experienced a unique situation of prosperity, with fast economic growth and systematic productivity gains combined with low unemployment and continuous wage growth. The intense social conflicts of previous decades led to the creation of a broad network of worker protection, characterized by the expansion in coverage and value of unemployment insurance, introduction of family allowances, indexation of pensions, and a set of measures that resulted in a significant change in the social policies (Glyn et al., 1990; Espin-Andersen, 1989). Low unemployment and strengthening of the Welfare State increased the bargaining power of workers in wage negotiations. Naturally, the achievements of the working class must be understood in the geopolitical context of the Cold War, in which it was necessary to improve the social and economic performance of capitalism in Western countries (Marglin, 1990; Korpi, 2002; Serrano 2004). In sum, the period of the Golden Age was characterized by a distributive pact that allowed the growth of real wages to accompany the rate of growth of labor productivity. The distributive dynamic during the 1950s and 1960s impacted the behavior of inflation – which was quite different from the one that had prevailed in the preceding decades. The so-called creeping inflation was characterized by moderate and persistent inflation rates, in a context of low levels of unemployment as well as high rates of growth. Notably, the inflationary process of advanced capitalist countries showed a stable dynamic, with no tendency to accelerate, while the Golden Age distributive arrangement remained relatively strong. However, the intensification of the distributive conflict, reflected in the increase of wage demands, raised the level of inflation rates in these countries by the end of the 1960’s. Finally, in the early 1970s, the liberation of exchange rate fluctuations, oil price shocks, and workers' resistance to absorbing the impact of the change in terms of exchange and currency devaluations - seeking to avoid the reduction of real wages - led to a change in the level of inflation rates, also causing deeper transformations in the dynamics of distributive conflict (Kaldor, 1976, Serrano, 2004, Korpi, 2002).

Creeping inflation was associated with wage inflation, at a time when the high bargaining power of workers ensured periodic readjustments of nominal wages. The growth of the nominal wage rate above the average rate of change in productivity implied the transfer of higher wage costs to prices, establishing the main cause of inflation in the period. In accordance with this process, the period of creeping inflation was characterized in different countries by a trajectory of real wages growth that was advancing fast and at approximately the same pace of productivity growth (Turnner and Jackson, 1970; Eatwell, Llewellyn and Tarling, 1974, Glyn et al., 1990).

Under those historical and economic conditions, alternatives of policy were debated aiming to deal with inflation without harming employment and growth. One of the outstanding contributions was the Scandinavian model of inflation. This model of inflation was created to address the trend of prices, wages and distribution in small (that is, price-taker) open economies under fixed exchange rate regime – as established after the Second World War by the Bretton Woods agreement. Developed initially by Odd Aukrust, the so-called Norwegian model resulted from research developed at the Central Bureau of Statistics of Norway during the 1960s and was first applied in Norway in 1966 (Aukrust, 1977). A second version of the model, which followed basically the same structure of the original, was elaborated a few years later by a group of Swedish authors linked to the workers’ and employers’ organizations that centralized the wage bargaining system: Gösta Edgren, Karl-Olof Faxén and Clas-Erik Odhner, representing the Swedish Central Organization of Salaried Employees (TCO), the Swedish Employers' Confederation (SAF) and the Swedish Confederation of Trade Unions (LO).

Under conditions of high degree of centralization of wage bargaining and with the economy close to full employment, economists aimed to contribute theoretically to the political process involving wage bargaining given – what they considered – a biding external constraint. In small open economies under fixed exchange regimes, the dynamics of international prices influences the path followed by prices, and given the dynamic of nominal wages and profits, their real values. Several prices of these economies are determined in the international markets, given that the country needs to keep its exports’ competitiveness there is little room to transfer an increase in domestic costs (such as wages) to final prices. An "excessive" wage growth could reduce the profitability of economic activities producing tradable goods, jeopardizing, as we mentioned before, the international competitiveness of these activities. In this case, problems in the trade balance would arise, and consequently the Balance of Payments would deteriorate more and more, until it limits the rate of economic growth. This issue is particularly relevant in the case of economies with a high degree of commercial openness such as the Scandinavian ones.

Thus, the Scandinavian model of inflation derives from the need to understand the relationship between the path of wages and prices during the Golden Age, when nations adopted fixed exchange rate regimes and there was a political and social pressure for the increase of real wages. Since international factors strongly influence prices and distribution in open economies, we expect that the discussion about Scandinavian model contributes to understanding inflation and distributive conflict in open economies in general, following a non-orthodox, cost push, approach.

The model starts from a fundamental distinction between economic activities that makes possible dividing the economy in two sectors: competitive sector – exposed to foreign competition – and sheltered sector – composed by activities that are not exposed to foreign competition, due to its own nature or to commercial protection. Firms from sheltered sector act in a competitive environment restricted to domestic economy. Sectors also differ in the dynamics of productivity, since it is assumed exposed sector is the leader in the growth of productivity. Besides that, the model assumes the economy is price taker at the international market, i.e., prices in the competitive sector are exogenous and are not affected by domestic variables. Moreover, the economy operates under fixed exchange rate regime (which was the common situation in the period). Hence, prices in competitive sector in domestic currency are directly determined by two exogenous variables: prices in international markets and exchange rate.

The main conclusion obtained is an inflation rate totally determined by foreign inflation and by a structural component. The latter reflects the differential between gains in productivity in the competitive sector and in the sheltered sector, weighted by the participation of the sheltered sector in the product. The growth rate of productivity in the competitive sector generates an inflationary pressure when it causes wage changes in the economy. On the other hand, the growth rate of productivity in the sheltered sector has the opposite effect, since it reduces the impacts of wage changes over costs and prices of the sheltered sector. Therefore, the bigger the productivity differential between the two sectors, the higher will be the inflation rate of the economy. Since inflation is explained by “structural” variables, authors claimed that the model provided a structural explanation to inflation (Aukrust, 1977; Edgren, Faxén and Odhner, 1973; Frisch, 1977).

From those results, authors draw some conclusions about anti-inflation policies. If the exchange rate is keep fixed, authorities are not able to influence long run trend of domestic inflation (Aukrust, 1977). As economic policy has no effect on international prices, it in incapable of affecting prices in competitive sector. Nor it impacts the productivity differential between competitive and sheltered sectors. In this case, policies that control aggregate demand should not be used for reducing inflation, since it would have no effective results, but could harm employment and growth (Aukrust, 1977; Edgren, Faxén and Odhner, 1969).

The approach proposed by the Scandinavian model should not be qualified as theory of inflation, as it is recognized by its proponents (Aukrust, 1977; Edgren, Faxén and Odhner, 1973; Frisch, 1977). The model tries to describe how foreign inflationary shocks propagate across the domestic economic activity, affecting inflation and income trends, without explaining theoretically the cause of international inflation. Nevertheless, the explanation for the rise in prices through the channel of wage costs to prices brings the model close to cost-inflation approach. Consequently, the model is compatible and contributes to the analysis of inflationary process within the cost-inflation and distributive conflict tradition (Okishio, 1977; Rowthorn, 1977; Serrano, 1993; Stirati, 2001; Lavoie, 2014). In this regard, it can be characterized as a wage inflation model, in which nominal wages growth is superior to the average productivity growth, in the case of small open economy, with different sectoral productivity rates of growth.

In the article, we provide an alternative closure to the model inspired by the contributions of Pivetti (1991), Serrano (1993, 2010) and Stirati (2001) to the comprehension of the inflationary process and the underlying dynamics of distributive conflict, in line with the Sraffian approach. According to this view, the basic distributive variables are determined in nominal values (nominal wages, nominal interest rate or nominal profit margin), influenced by political and institutional factors. Inflation emerges from the incompatibility of distributive variables with the aggregated level of prices and the structure of relative prices in the economy. Distributive variables in real terms (real wages, real interest rate or real profit margin) consist in an ex post result obtained under certain inflationary development, which reflects some accommodation of distribution.

Following this approach, we can consider that prices are determined by a nominal mark-up on labor unit cost production costs. The relevant costs for the determination of prices under competition are the historical costs of capital. However, in our first analytical exercise, we consider labor as the only cost of production of both sectors, for the sake of simplicity. On the other hand, the real mark-up expresses the ratio of prices to replacement costs of capital, which is coherent to income distribution observed at the end of period (and with the profit rate, as it is normally understood). Under a situation of persistent inflation, nominal mark-up and real mark-up differ as production costs increase between the period in which production begins and the period in which capital replacement occurs. If we reduce production costs to labor and assume that productivity does not change, the real mark-up obtained ex post will be equal to the ratio between the nominal mark-up to the wage growth rate (Serrano, 2010). Allowing for distribution to change, we can explain divergent paths (as the ones observed by Scandinavian economies) and also achieve a long-term solution under a different pattern of distribution of income between profits and wages.

Nevertheless, in spite of analytical limitations and structural-historical specificities that conditioned the Scandinavian model, it represented a pioneering formulation for the theory of cost inflation in the context of open economies. Thus, we expect that the extensive analysis carried out in this article contributes to the study of the interaction between inflation, distribution and international prices. Of course, future research should take into account the behavior of the exchange rate, a variable that in many countries has proved to be very relevant to distributive conflict.

This article aims to critically analyze the Scandinavian model of inflation, considering the recent contributions of Sraffian approach to inflation theory. Besides a short introduction describing relevant historical conditions, the article is organized in four other sections followed by closing remarks. Section 2 presents the Scandinavian model in its original version and policy implications pointed by its authors. The following section discusses the limitations of the Scandinavian model analysis of distribution. By its turn, section 4 and 5 present alternative approaches to solve the model, flexing some of its hypothesis and obtaining new conclusions. Finally, we briefly discuss the main results in the conclusion section.

During the Golden Age, the advanced capitalist economies experienced a unique situation of prosperity, with fast economic growth and systematic productivity gains combined with low unemployment and continuous wage growth. The intense social conflicts of previous decades led to the creation of a broad network of worker protection, characterized by the expansion in coverage and value of unemployment insurance, introduction of family allowances, indexation of pensions, and a set of measures that resulted in a significant change in the social policies (Glyn et al., 1990; Espin-Andersen, 1989). Low unemployment and strengthening of the Welfare State increased the bargaining power of workers in wage negotiations. Naturally, the achievements of the working class must be understood in the geopolitical context of the Cold War, in which it was necessary to improve the social and economic performance of capitalism in Western countries (Marglin, 1990; Korpi, 2002; Serrano 2004). In sum, the period of the Golden Age was characterized by a distributive pact that allowed the growth of real wages to accompany the rate of growth of labor productivity. The distributive dynamic during the 1950s and 1960s impacted the behavior of inflation – which was quite different from the one that had prevailed in the preceding decades. The so-called creeping inflation was characterized by moderate and persistent inflation rates, in a context of low levels of unemployment as well as high rates of growth. Notably, the inflationary process of advanced capitalist countries showed a stable dynamic, with no tendency to accelerate, while the Golden Age distributive arrangement remained relatively strong. However, the intensification of the distributive conflict, reflected in the increase of wage demands, raised the level of inflation rates in these countries by the end of the 1960’s. Finally, in the early 1970s, the liberation of exchange rate fluctuations, oil price shocks, and workers' resistance to absorbing the impact of the change in terms of exchange and currency devaluations - seeking to avoid the reduction of real wages - led to a change in the level of inflation rates, also causing deeper transformations in the dynamics of distributive conflict (Kaldor, 1976, Serrano, 2004, Korpi, 2002).

Creeping inflation was associated with wage inflation, at a time when the high bargaining power of workers ensured periodic readjustments of nominal wages. The growth of the nominal wage rate above the average rate of change in productivity implied the transfer of higher wage costs to prices, establishing the main cause of inflation in the period. In accordance with this process, the period of creeping inflation was characterized in different countries by a trajectory of real wages growth that was advancing fast and at approximately the same pace of productivity growth (Turnner and Jackson, 1970; Eatwell, Llewellyn and Tarling, 1974, Glyn et al., 1990).

Under those historical and economic conditions, alternatives of policy were debated aiming to deal with inflation without harming employment and growth. One of the outstanding contributions was the Scandinavian model of inflation. This model of inflation was created to address the trend of prices, wages and distribution in small (that is, price-taker) open economies under fixed exchange rate regime – as established after the Second World War by the Bretton Woods agreement. Developed initially by Odd Aukrust, the so-called Norwegian model resulted from research developed at the Central Bureau of Statistics of Norway during the 1960s and was first applied in Norway in 1966 (Aukrust, 1977). A second version of the model, which followed basically the same structure of the original, was elaborated a few years later by a group of Swedish authors linked to the workers’ and employers’ organizations that centralized the wage bargaining system: Gösta Edgren, Karl-Olof Faxén and Clas-Erik Odhner, representing the Swedish Central Organization of Salaried Employees (TCO), the Swedish Employers' Confederation (SAF) and the Swedish Confederation of Trade Unions (LO).

Under conditions of high degree of centralization of wage bargaining and with the economy close to full employment, economists aimed to contribute theoretically to the political process involving wage bargaining given – what they considered – a biding external constraint. In small open economies under fixed exchange regimes, the dynamics of international prices influences the path followed by prices, and given the dynamic of nominal wages and profits, their real values. Several prices of these economies are determined in the international markets, given that the country needs to keep its exports’ competitiveness there is little room to transfer an increase in domestic costs (such as wages) to final prices. An "excessive" wage growth could reduce the profitability of economic activities producing tradable goods, jeopardizing, as we mentioned before, the international competitiveness of these activities. In this case, problems in the trade balance would arise, and consequently the Balance of Payments would deteriorate more and more, until it limits the rate of economic growth. This issue is particularly relevant in the case of economies with a high degree of commercial openness such as the Scandinavian ones.

Thus, the Scandinavian model of inflation derives from the need to understand the relationship between the path of wages and prices during the Golden Age, when nations adopted fixed exchange rate regimes and there was a political and social pressure for the increase of real wages. Since international factors strongly influence prices and distribution in open economies, we expect that the discussion about Scandinavian model contributes to understanding inflation and distributive conflict in open economies in general, following a non-orthodox, cost push, approach.

The model starts from a fundamental distinction between economic activities that makes possible dividing the economy in two sectors: competitive sector – exposed to foreign competition – and sheltered sector – composed by activities that are not exposed to foreign competition, due to its own nature or to commercial protection. Firms from sheltered sector act in a competitive environment restricted to domestic economy. Sectors also differ in the dynamics of productivity, since it is assumed exposed sector is the leader in the growth of productivity. Besides that, the model assumes the economy is price taker at the international market, i.e., prices in the competitive sector are exogenous and are not affected by domestic variables. Moreover, the economy operates under fixed exchange rate regime (which was the common situation in the period). Hence, prices in competitive sector in domestic currency are directly determined by two exogenous variables: prices in international markets and exchange rate.

The main conclusion obtained is an inflation rate totally determined by foreign inflation and by a structural component. The latter reflects the differential between gains in productivity in the competitive sector and in the sheltered sector, weighted by the participation of the sheltered sector in the product. The growth rate of productivity in the competitive sector generates an inflationary pressure when it causes wage changes in the economy. On the other hand, the growth rate of productivity in the sheltered sector has the opposite effect, since it reduces the impacts of wage changes over costs and prices of the sheltered sector. Therefore, the bigger the productivity differential between the two sectors, the higher will be the inflation rate of the economy. Since inflation is explained by “structural” variables, authors claimed that the model provided a structural explanation to inflation (Aukrust, 1977; Edgren, Faxén and Odhner, 1973; Frisch, 1977).

From those results, authors draw some conclusions about anti-inflation policies. If the exchange rate is keep fixed, authorities are not able to influence long run trend of domestic inflation (Aukrust, 1977). As economic policy has no effect on international prices, it in incapable of affecting prices in competitive sector. Nor it impacts the productivity differential between competitive and sheltered sectors. In this case, policies that control aggregate demand should not be used for reducing inflation, since it would have no effective results, but could harm employment and growth (Aukrust, 1977; Edgren, Faxén and Odhner, 1969).

The approach proposed by the Scandinavian model should not be qualified as theory of inflation, as it is recognized by its proponents (Aukrust, 1977; Edgren, Faxén and Odhner, 1973; Frisch, 1977). The model tries to describe how foreign inflationary shocks propagate across the domestic economic activity, affecting inflation and income trends, without explaining theoretically the cause of international inflation. Nevertheless, the explanation for the rise in prices through the channel of wage costs to prices brings the model close to cost-inflation approach. Consequently, the model is compatible and contributes to the analysis of inflationary process within the cost-inflation and distributive conflict tradition (Okishio, 1977; Rowthorn, 1977; Serrano, 1993; Stirati, 2001; Lavoie, 2014). In this regard, it can be characterized as a wage inflation model, in which nominal wages growth is superior to the average productivity growth, in the case of small open economy, with different sectoral productivity rates of growth.

In the article, we provide an alternative closure to the model inspired by the contributions of Pivetti (1991), Serrano (1993, 2010) and Stirati (2001) to the comprehension of the inflationary process and the underlying dynamics of distributive conflict, in line with the Sraffian approach. According to this view, the basic distributive variables are determined in nominal values (nominal wages, nominal interest rate or nominal profit margin), influenced by political and institutional factors. Inflation emerges from the incompatibility of distributive variables with the aggregated level of prices and the structure of relative prices in the economy. Distributive variables in real terms (real wages, real interest rate or real profit margin) consist in an ex post result obtained under certain inflationary development, which reflects some accommodation of distribution.

Following this approach, we can consider that prices are determined by a nominal mark-up on labor unit cost production costs. The relevant costs for the determination of prices under competition are the historical costs of capital. However, in our first analytical exercise, we consider labor as the only cost of production of both sectors, for the sake of simplicity. On the other hand, the real mark-up expresses the ratio of prices to replacement costs of capital, which is coherent to income distribution observed at the end of period (and with the profit rate, as it is normally understood). Under a situation of persistent inflation, nominal mark-up and real mark-up differ as production costs increase between the period in which production begins and the period in which capital replacement occurs. If we reduce production costs to labor and assume that productivity does not change, the real mark-up obtained ex post will be equal to the ratio between the nominal mark-up to the wage growth rate (Serrano, 2010). Allowing for distribution to change, we can explain divergent paths (as the ones observed by Scandinavian economies) and also achieve a long-term solution under a different pattern of distribution of income between profits and wages.

Nevertheless, in spite of analytical limitations and structural-historical specificities that conditioned the Scandinavian model, it represented a pioneering formulation for the theory of cost inflation in the context of open economies. Thus, we expect that the extensive analysis carried out in this article contributes to the study of the interaction between inflation, distribution and international prices. Of course, future research should take into account the behavior of the exchange rate, a variable that in many countries has proved to be very relevant to distributive conflict.