Value and quality creation: natural resources, industrialization and standards of living in peru 1950 to 1997

Peru”s recent economic policy, like that of most Latin American countries! has followed the paradigm of the socalled “Washington Consensus”. Such paradigm precludes the implementation of “strategic” industrial policies as well as the active and deliberate construction of competitive advantages through measures that foster certain sectors or activities. “Washington-Consensus” thinkers hold that the “magic of the market” and its indiscriminate opening will allow countries to acquire the necessary long-term external competitiveness, promote economic growth and enhance standards of living, regardless of the country”s productive specialization.

trade in specific products has impacted economic growth and standards of living in Peru. It seeks to determine whether the prevalence of producing or trading goods from specific branches or sectors, whether natural resources, industrial or services, have either a positive or negative impact on the people's standard of living and wellbeing. To this end, we analyze Peruvian data for the last 50 years, paying special attention to the expansion and recession cycles, and to industrialization policies implemented in Peru's economy since the end of the 1950s.
Our two core hypotheses are that: i) the standard of living is inversely related to the level of "primary" activities. In other words, the general well-being of the * Una versión preliminar de este trabajo fue presentada a la conferencia internacional "Globalization and Marginalization in the 1990s: An Evolutionary and Activity Specific Perspective", organizada por SUM-Center for Development and the Environment, University of Oslo & Norsk Investorforum, en Oslo entre el 5 y el 6 de setiembre de 1998. Traducido del original en español por Sidney Evans, con la supervisión del primero de los autores, el profesor Santiago Roca.
esan-cuadernos de difusión 70 population will decline as national economic activity is increasingly directed towards the development of raw materials or extractive activities; and ii) the economy will become increasingly fragile if, as a result of productive specialization, there is an "uncoupling of quality and value" between the country's exports and imports. In other words, the country will become increasingly fragile if it trades an increasing amount of simple goods in exchange for the same or a smaller amount of more elaborate foreign goods.
In the first section we present a brief overview of the theoretical framework that explains how economic activities differentiate one from the other and why countries that specialize in producing and trading goods of "lower quality or value" eventually also lower their own living standards and well-being. The second section shows quantitative data on standards of living and the main types of activities that have prevailed in Peru over the last half-century. The main findings concerning the relationship between primarization, industrialization and standards of living appear in the third section, while the fourth section provides a first evaluation of the "uncoupling of quality and value" between Peruvian exports and imports. Lastly, we give some recommendations for Peruvians to benefit from the new era of globalization and trade.

THEORETICAL CONSIDERATIONS
Most classical literature about economic growth holds that economic activities do not matter and that economic growth depends on the abundance and best use of capital, labor or natural resources, as well as technology, infrastructure, free trade, government's efficiency, savings and investment, education, individual effort, driving force of the people, etc. However, little or even no attention has been paid to economists who underscore the type of products a country turns out as a main factor for economic growth. Reinert (1993Reinert ( , 1995 has explored more than 500 years of economic history to show that economic growth and standards of living depend on the type of activities performed by nations. Graham (1923) presents one of the soundest arguments that specialization in one type of activity or another is crucial for growth and commerce. Based on a simple, before-andafter numerical trade example, Graham showed that the standard of living and level of income for two countries who specialize in production and engage in trade on the basis of their comparative advantages will improve for such two nations and the world at large if and only if both countries can produce at similar returns.
If a country specializes in producing goods with increasing returns while the other country specializes in producing goods with diminishing returns, the world at large will also increase its income but revenues in the country specialized in goods that are produced at diminishing returns will decrease, while revenues in the country with growing increasing returns will rise. This means one country will be wealthier and the other poorer although the world will become wealthier as a whole 2 . Graham, and many previous Value and quality creation thinkers in past centuries like Serra (1613), Colbert (1651), Rosher (1882), Raymond (1820), hold that raw materials and natural resources intrinsically provide diminishing returns while manufactured products show increasing returns. In Annex 1 we have further elaborated on Graham's initial exercise to show that trade between countries with similar returns favors both, but that if one has growing or constant returns and the other has diminishing returns, the first one will prosper to the disadvantage of the second one, where standards of living and revenue will fall despite a larger world product.
Another argument underscoring the fact that the type of productive activity has an influence on economic growth, though from a different vantage point, was advanced in the 1950s and 1960s by authors including Hirschman (1961), Prebish (1970), Singer (1981), Seers (1975) and Myrdal (1963), to mention just a few. Prebish holds that countries specializing in raw materials and natural resources are harmed because the prices of raw materials grow relatively less rapidly than prices of manufactured products. Hirschman holds that agriculture lacks the upstream and downstream linkages or the complex division of labor that characterizes manufacturing. Myrdal mentions the "cumulative causation" present in manufacturing, but not found in natural resources development. Others point to higher income elasticity and increased growth of demand for manufacturing goods compared to primary products.
More recently, in the 1990s, Matsuyama (1992) and others pointed to the fact that manufacturing shows positive growth effects that are absent in agriculture and stem from higher, "induced learning." In other words, there are a number of learning externalities that neither agriculture nor the service sector can provide. Sachs and Warmer (1995) made an empirical, comparative world survey showing that countries richly endowed with natural resources grow less than countries specializing in tradable manufactured products. However, they fail to explain the reasons underlying those differences and then argue that free trade is beneficial for all the involved parties.
Within the theory of trade, Krugman (1991), and Krugman and Obsteld (1995) hold that an economic activity may be better than another only if there are market imperfections that include positive externalities originating in technological innovation or the existence of rents in highly concentrated oligopolistic industries 3 . competition will prevail. From this theory, we can infer that international commerce will bring benefits to all nations that trade under such conditions. In the 1980s, Krugman focused on part of Graham's work to redesign the whole theory of trade under the assumption of increasing returns and verified that in this case trade would also benefit all countries. However, he failed to consider, as correctly pointed by Reinert (1996), the possibility that one country would show growing returns and the other decreasing ones, in which case the former becomes richer and the latter poorer, as demonstrated by Graham in 1923. Reinert (1993,1996) takes a more dynamic, encompassing and integrating viewpoint of economic history that approaches and complements theories about competitiveness and corporate strategies by authors including Porter (1990), Marrus (1984), David (1986) and others. Reinert argues that economic activities are different from a qualitative vantagepoint, and that they determine economic growth and income disparities among nations. In this respect, Reinert proposes a "quality index" for economic activities whereby countries that grow are those that focus on producing "high quality" goods while nations that produce "low quality" grow less or move backwards.
"High quality goods" typically feature increasing returns and are performed under conditions of imperfect competition with steeply sloping learning curves, rapid technological changes, large-scale R&D and investment, high growth and income demand elasticity rates, economies of scale, numerous linkages, a high and complex division of labor, a significant need to "learn by doing," high industrial concentration, imperfect and at the same time extremely dynamic information flows, high barriers to market entry and exit, high salaries, etc. These are all industries that closely follow the assumptions of Schumpeter's theory of imperfect though dynamic and changing competition.
Low quality goods are characterized by decreasing returns, perfect competition, flat learning curves, slow technological development, small R&D investment, low growth and demand income elasticity, falling economies of scale, few linkages, scant division of labor, reduced requirement for institutional learning, a fragmented industry, perfect information, low entrance and exit barriers, low wages, etc. In other words, all those activities that closely reflect the traditional assumptions of the Neoclassical theory of production, trade and growth.
According to Reinert (1996), trade among industrial and non-industrial nations is characterized by asymmetrical exchanges between industries that on the one hand, feature large economies of scale and evolve under conditions of imperfect competition and, on the other industries characterized by diminishing returns and perfect competition. Moreover, he holds that present, industrial economies historically chose to follow an active and deliberate path to "high quality goods production," by enforcing highly successful "industrial and commercial strategies" which they, having now achieved productive excellence, no longer recommend as a recipe for growth to developing nations. duce a marginal social benefit or will generate positive externalities that spread to the rest of the economy. In such cases, there should be in theory some subsidy mechanism to allow the high-technology industries (v.g. biotechnology, electronics, space industry, etc.) to capture some of those social benefits and thus further foster its own development (Grossman, 1991). In the case of oligopolistic industries, given the small number of companies that trade on a world scale, windfall profits are a logical consequence. Under such circumstances, a country may likely subsidize its own company so as to push other countries' companies out of the market. This is commonly called strategic commercial policy. However, such policy may unleash a trade war if all the other countries react likewise (Brunder, 1991). Despite these arguments, the fundamental implications for economic policy derived from such proposals are still valid. Thus, for instance, Value and quality creation 2. STANDARD OF LIVING AND TYPE OF ECONOMIC ACTIVITY. PERU 1950PERU -1997 In the hypothesis that growth and standards of living depend on a country's productive specialization, this section analyzes the evolution of standards of living and the main economic activities that have characterized Peru's last half century.

Legal and institutional aspects. Industrial and labor policies
Both the type of economic activity and the standard of living of any economy are influenced not just by economic and market cycles, but also by industrial, labor and income policies and by institutional factors and diverse cultural elements. We do not aim here at identifying each and every one of these factors, but it is necessary to take into account the wider trade, industrial and income policy periods that Peru has evolved in the last 50 years.
Since 1959, Peru's productive activities were, to varying extents influenced by the Industrial Promotion Law (Law No. 13270) which introduced substantial incentives to industrial investment, particularly those in basic and decentralized activities. Some of the schemes included in this law underscored tax incentives to reinvestment and growing effective protection for the manufacturing industry by reducing tariffs to imports of capital goods and inputs, while more heavily taxing imports of consumer goods (see Ferrari, 1992;Portocarrero and Nunura, 1984).
The first law enacted in 1959 was followed by the 1970 Industrial Promotion Act (Law No. 18350) and the corresponding 1981 and 1986 bills that slightly modified the initial regulations without canceling the crucial role afforded to incentive policies, and to the State's role as regulator, planner and even producer, as was characteristic of the industrial policy introduced in 1970 when Stateowned companies started to expand.
The concept of priority industries was introduced in 1970 to underscore and foster through tax, tariff, credit and administrative schemes, the development of basic industries including cement, paper, basic chemicals, steel, fertilizers and oil refining, all of which were reserved exclusively to the State. In 1981 and 1986, although the existing industrial policy was largely preserved, the State's monopoly and the definition of priority industries were canceled, as well as tax exemptions for reinvested profits. Occasionally, also the option to repatriate profits and royalties was suspended.
Generally speaking, from the 1960s to the 1980s, industrialization policies favored permanent protection for all types of industries devoted to producing finalgoods for the local market where there was little internal competition. There existed no "learning" processes or linkages with foreign countries, technology was imported, and the policy as a whole was enforced with a static planning vision it is held that no conclusive empirical evidence allows to hold that markets will necessarily fail. Additionally, even if this were true, the criteria selected to foster new activities are not clear, nor is there sufficient information to evaluate and determine what industries should be promoted. Lastly, it is also held that this policy is not free from various pressures which would, in the best of cases, hamper making final decisions (Grossman, 1991).

esan-cuadernos de difusión 74
of the world and business. This view sought to create an industry to assemble imported parts and components without paying attention to education, creativity and training. It thus had scarce possibilities of linking up with a broader market. Income and job allocation in industry were ruled by static rent and profit seeking attitude, and by a culture based on nepotism rather than merit. Government assumed an oversized entrepreneurial role that lacked synergies and eventually excluded and displaced domestic and foreign private investment.
On the side of revenue policies, the first half of the 1970s saw the emergence of measures and reforms that initially fostered wage and salary growth and which were related to the land and company reforms. Likewise, a number of labor regulations enacted in that period made lay-off more difficult while giving workers job tenure and strengthening unions.
In 1990's all, the previous industrial, labor and income policy schemes were cancelled. The new government policy suspended the main tax exemptions, reduced tariff structure, and lifted price controls while financial, exchange and trade regulations were liberalized. Job tenure was also eliminated and the whole labor legislation was made more flexible, thus reducing union power. A radical government downsizing program through privatization and/or the sale of Stateowned assets was introduced and, all restrictions to the inflow or outflow or foreign private capital were lifted.
As we shall see below, the relative importance of the manufacturing sector grew through the enforcement of industrial policies only until the mid-1970s, and was later reduced in the wake of earlier stabilization programs. To the extent that industrial development was based on developing the domestic market through industrialization focusing on import substitution, the enforcement of adjustment programs inevitably led to receding national industrialization.
Concurrently, standards of living and income levels fell as a consequence of booming population growth and lower factor productivity.

GDP evolution and main economic activities
Peru's economy grew at an annual 3,3% average rate 4 from 1950 to 1997 but its annual performance was very irregular and less than satisfactory (see Figure 1). In the initial 20 years from 1950 to 1970, the annual average growth rate reached 5% with only two years of stagnation and recession in 1958 (-0,6%) and 1968 (0,4%).
Between 1970 and 1990, annual growth was 1,7% including several periods of strong economic contraction due not only to weather difficulties such as the 1983 El Niño weather anomaly (-12,6%) and drought in 1992 (-1,4%), but also as a result of adjustment and stabilization programs in 1976-78, 1988-89 and 1990. Finally, in the last 7 years, the economy has grown at an annual 5,7% rate despite which production in 1997 was only 13% larger than in 1987.

Value and quality creation
A breakdown of GDP by period and economic sector in the 1950s and 1960s reveals that fisheries activities led growth at an annual average rate of 19% and 9,3% respectively, where energy (electricity, gas and water) grew 8,8% and 7,4%, and manufacturing expanded 7,1% and 5,5%, for the respective years. Such levels of growth gave these productive sectors a significatively larger share in global GDP in the first two decades of the periods under consideration.

esan-cuadernos de difusión 76
to the benefit of agriculture and construction with less importance attributed to mining and government services.
Generally, the industrial policy enforced since the end of the 1950s led to the development of the manufacturing industry that increased its share from 19% in 1950 to a maximum 25,5% in 1976. However, together with poor management of the agricultural companies created by the state, this industrial policy led to the relative fall of agriculture, which dropped from 23,7% in 1950 to 9,9% of GDP in 1980.
In the 1980s and 1990s, manufacturing took a step backwards compared to the 1970s with a clear trend towards deindustrialization. The reasons can be found in the effects of the recession provoked by stabilization programs and by the suspension since 1990 of various industrial promotion and protection schemes and incentives that had been in place since 1959. Evolution in most other sectors was basically influenced by exogenous factors, be they foreign (as in mining), weather (fisheries), population (electricity, gas and water) or relating to the expansion of government expenditure (as in construction and government services).
An easier way to classify productive specialization in Peru over the past half century is by dividing GDP into four large economic groups or sectors: (i) extractive or primary activities (agriculture, fisheries and mining); (ii) basic transformation or infrastructure (construction); (iii) intermediate or industrial transformation (manufacturing); and (iv) services (home rentals, government, electricity, gas and water, commerce, services and others).
Although this standardized classification is rather broad and does not accurately reflect the "quality index for economic activities" proposed in the theoretical framework, we still do not have a methodology that will allow to classify economic activities by use-intensity and technological upgrading capabilities, nor from the viewpoint of their relationship Despite these constraints, we can assume that on average, natural resources activities are extractive and create goods with diminishing returns in perfectly competitive markets with low salaries. Infrastructure and manufacturing activities are processes with increasing returns, operating in imperfect markets, with larger technology investments, higher salaries, etc. The service sector sits somewhere between the two 6 . Table 3 shows GDP structure according to the four suggested sectors. Clearly, in the first quarter century, the primary sector's participation decreased by a significant 13%, from 32% in 1950 to 19% in 1975. In the 80s and 90s, these sectors' importance increased again, to between 22% and 25%.
Manufacturing activities (including processing of primary resources such as fishmeal, frozen fish, sugar, non-ferrous metals and refined oil) also increased their share as a result of industrialization policies and reached a maximum 25% of GDP in 1975 to then fall again to 23,8% in 1980, 22,1% in 1990 and 22% in 1997. 5 More recently, the Organization for Economic Cooperation and Development (OECD, 1992) suggested a difference between supplier-driven activities, economy-of-scale intensive activities and those based on science. However, so far national statistics bureaus continue to use the International Industrial Uniform Classification presented in Tables 1 and 2 and proposed by the United Nations.
6 Obviously, this classification implies some over simplification that must be corrected in later work. For instance, the garment and electronic industries are not comparable from a qualitative viewpoint for their effect on technological development, use of R &D, level of salaries, economies of scale, externalities, etc. Likewise, there is a substantial difference between extensive agriculture, and capital and technologyintensive mining or fisheries. Despite these constraints, this type of classification makes it possible to establish the differences between and quantify those activities that generally allow to reach increasing returns compared to those where only diminishing returns are possible if we follow Ricardo's terminology. Throughout the period under consideration, construction never increased its share of GDP above 7% except in 1955 when it reached 7,9% and in recent years when, propelled by government expenditures and expanded home building in the private sector, it reached almost 9% of GDP.
Lastly, services increased their relative share from 43% in the 1950 to 47% in the 1990s.

Standard of living: per capita product, per capita consumption and remuneration
A simple way to classify levels of income and standards of living is through indicators like per capita product and consumption, and wages (white-collar income) and salaries (blue-collar income). Although it is true these indicators may hide inequality and income distribution disparities, they do constitute reasonable criteria for this paper.
Taking into consideration population has grown at an annual average rate of 2,6%, GDP per capita from 1950 to 1997 grew annually at a 0,7% rate, and private per capita consumption grew by 0,6%. At these rates for income per capita and population's purchasing power to double, we would have to wait at least one hundred years.
Despite such overall poor performance, we can observe rapid growth in the first 25 years of the period under study, which led GDP and consumption per capita in the mid-1970s to rise 80% above the corresponding 1950 figures. However, in subsequent years there was a notorious falling trend so that per capita income in 1997, although 69% higher than in the 1950s, was 11% lower than the historical record reached in 1981. Moreover, current per capita private consumption is 52% higher that of the 1950s but 17% below the historical record achieved in 1975 (see Figure 2).  Table 4 shows that GDP and per capita consumption growth have not evolved in parallel. Quite the contrary, there have been alternating periods where income growth exceeded consumption growth and, conversely, in other periods consumption exceeded the growth of income. Thus, in the 1950s, 1970s and 1990s per capita GDP growth exceeded the growth of private consumption while it was lower in the 1960s and 80s.
Between 1960 and 1997, real wages (white-collar) and salaries (blue-collar) fell at an average annual -4,2% and -3,7% rates, respectively, thus diminishing more steeply than per capita revenues and consumption. Workers' earnings were most severely hit during the 1980s when they fell between 9% and 10% a year so that earnings in 1990 were scarcely between 21% and 29% of those prevailing 1960 ( Figure 3).
Generally, living standards over the last half century measured either on the side of per capita private consumption or real wages and salaries were characterized by an upward trend until the mid-70s, and then, by a strong contraction with the introduction of economic stabilization programs. However, wages and salaries have suffered a stronger decline (-75% since 1973) than per capita consumption (-17% since 1975) showing those white and blue-collar wage and salary earners bore the brunt of economic adjustment.

AND STANDARD OF LIVING
This section analyzes the relationship between Peru's productive specialization categorized by product as proposed in Section 2.2, and the population's standard of living and income estimated in Section 2.3. More specifically, it determines whether higher relative development of primary goods, infrastructure, manufacturing or services is linked to higher living standards measured through wages and salaries or per capita private consumption.
A relationship between productive specialization and standard of living is proposed both for the long and short-term periods. In the long run economic growth factors most relevant are: increased productivity, economies of scale, technological innovation, labor specialization, capital stock increases, etc. In the short term macroeconomic fundamentals and imbalances are the major key factor.

The long term factor: Specialization in sectors with diminishing returns and poor technological development
By plotting the percent variation in living standards to the vertical axis and the changes in primary activities as a percent- 1960 1970 1980 1990 White collar wages Blue collar wages -50% -75%

Figure 3: REAL REMUNERATION (∆% ACCUMULATED) Value and quality creation
age of GDP to the horizontal axis (figures 4, 5 and 6) we observe an inverse (or negative) relationship between the relative importance of primary or extractive activities (such as agriculture, fisheries and mining) and private per capita consumption or real earnings. In other words, higher participation of primary activities leads to lower private per capita consumption, salaries and wages.

Figures 7, 8 and 9
show a similar but opposite relationship for manufacturing. As industrialization increases, higher per capita consumption and earnings (salaries and wages) are observed, meaning that increases in the industrialization index would imply higher standards of living.
A simple regression analysis by least squares (see Annex 2) shows that for every incremental percent point of extractive activities, private consumption falls by approximately -2,6% while white-collar wages fall by -5,4% and blue-collar salaries by -7,4%. In other words, the adverse impact is greater on earnings as a whole than on consumption.
On the other hand, an extra percent point in the share of manufacturing activities would increase per capita consumption by 4,2%, white-collar real wages by 10,6% and blue-collar real salaries by 15,5%. This means that manufacturing specialization not only increases standards of living but has a proportionally larger impact on blue-collar salaries, thus leading to a positive effect on income distribution.
In the construction industry, the impact on the various standard of living indicators would also be positive although the respective parameters are substantially smaller than those for manufacturing and are of little statistical significance 7 . Lastly, impact in the service sector would be close to zero with little statistical significance in either private consumption or earnings.
If economic primarization has the long-term effect of reducing the population's standard of living, why has there been such a long term insistence on producing primary goods? Two fundamental explanations are in order. The first, presented below, deals with the way the country participates in the world economy. The second relates to macroeconomic imbalances and will be analyzed later when we deal with the short-term factors affecting industrialization.
Peru's conventional exports and a large portion of its non-conventional exports are resource-based. Approximately 80% of total exports are related to agricultural, mining and fisheries industries while only 20% are related to manufacturing (Annex 3). This type of participation in international trade based on the use of natural resources has led the growth of exports in the last 50 years. It has also led to increasing economic primarization, a fact confirmed by a positive correlation (r2=0,36) between primarization and exports as a percentage of GDP ( Figure  10). Likewise, there is a positive correlation between real exchange rate and primarization, where r2=0,29 ( Figure 11). This shows that devaluation in real terms would increase the relative importance of primary activities. 7 Impact is statistically significant only at 5% per capita consumption but not for earnings, where statistical significance starts at 15%.

esan-cuadernos de difusión 84
Why would a higher real exchange rate lead to the increased relative participation of primary activities? Three reasons explain this result: 1) the physical quantity of raw material exports would increase as higher real exchange rate increases their profitability; 2) the increase of raw materials exports is higher than the expansion of manufacturing product (for either the domestic or external market) which would also grow as the real exchange rate increases; and, 3) indirectly, primarization grows as manufacturing output is reduced given the eventual negative impacts of higher real exchange rates on domestic demand. In these three instances real term devaluation would reduce the manufacturing to GDP ratio, which would in turn translate into higher participation of primary activities.
This happens because the country's industrialization process revolved around the domestic market and not the external market. In other words, it was not supported by the active and efficient promotion of manufacturing exports. If industrialization and the implicit trade policy were export-oriented, the slope of the curve in Figure 11 would be steeper and may even become negative because devaluation would eventually reduce eco-nomic primarization. For this to happen, however, the whole national industrial and trade strategy would have to be redefined, compared to recent decades.

Short term factors: Macroeconomic imbalances
A second reason explaining why it has not been possible to revert economic primarization relates to short term macroeconomic imbalances.
Although we showed that in the long term industrial development would lead to increased standards of living and revenues (earnings and per capita consumption), there is a fundamental difficulty in maintaining basic macroeconomic balances. This difficulty determines swings between industrialization and primarization or, seen from the other side, between primarization and de-primarization of economic activities (see Figure 12) 8 . If we separate in three periods the relationship between per capita consumption and primary activities to GDP, 1971GDP, -1978GDP, , 1979GDP, -1984GDP, and, 1985GDP, -1990, we may see that within each period there are years when primary activities contract and therefore industrialization and standards of living increase, followed by other years where the reverse process occurs, that is the relative importance of extractive activities grows again, with the consequent reduction in living standards.
For example, from 1971 to 1975 ( Figure  13), the rate of primary activities to GDP fell from 23,4% to 19,2% while per capita consumption increased by 6,9% (from 136 to 146, in 1979 soles). However, in the three years that followed, primary activities grew again to 23,7% while consumption dropped by 8,7%, for a standard of living that was 7% lower than at the beginning of this stage in 1971 -despite quite similar primarization levels.   In the second case (Figure 14), we observe that from 1979 to 1981, participation of the primary sector dropped from 24,8% to 22,9% pushing consumption up by 5%. In subsequent three years, primary activities grew again to 24,7% while consumption fell by 12,7%, even below 1979 consumption levels. The effects of the 1983 El Niño weather phenomenon though could largely account for this fact.
Lastly, in Figure 15, our third example shows that primary activities as a percentage of GDP fell by 3,5% points (from 21,5% to 21,6% between 1985 and 1987) while per capita consumption increased strongly by 20%. However, in the four years after that, the importance of extractive activities rose again to 24,7% of total GDP while per capita consumption plummeted by 29,4% or below its 1985 level when the process started.
In these three instances, such pendulum movement led the level of primarization almost to its point of departure. However, per capita consumption ultimately fell to a significantly lower level than its initial point. Consequently, these processes are extremely damaging to the standards of living of the population at large, because besides the pendulum swing from right to left, we can also notice a downward trend. By comparing the coordinates for 1979, 1985 and 1990, we see that the level of primary activities to GDP is roughly similar. Still, per capita consumption was significantly lower in 1990 than in 1985 and 1979. In other words, from the viewpoint of standards of living, the progress made in the 1980s was totally wasted.
The trend towards economic reprimarization started in 1976, 1982, and 1988, when the trade gap became unbearable and the country's foreign currency reserves did not suffice to sustain the imports of goods and services (See Table  6) 9 10 .
In the present decade (see Figure 16), the liberalization policy has translated into a slight decrease in the level of primary activities from 24% in 1991 to 22,5% in 1997, parallel to an increase in per capita consumption of 22,7% propelled by higher earnings and expanded consumer loans.
An important factor in the 90's leading to falling primary activities to GDP ratio is the strong growth of the construction industry to reconstruct the basic social infrastructure damaged by external factors such as weather and terrorism. Also influential was the opening of the economy 9 Insufficiency of foreign currency reserves can be partly accounted for by excess domestic demand which in the periods before the adjustment grew more than GDP. An important factor generating this growth of demand is obviously the increased private consumption, meaning that such swings are also explained by increased standards of living that create bottlenecks in the external sector. Of similar importance is the adverse impact of external shocks, such as falling terms of trade or higher international interest rates during the 1980s. Thus, for instance, a drop in terms of trade in 1975 (-31%), in 1981 and 1982 (-32%) and again in 1986 (-27%) also had a substantial impact on the trade balance and the amount of net foreign reserves which eventually sparked the stabilization programs. Value and quality creation and privatization that led to substantial investment growth, in particular foreign investment, in non-tradable activities such as commerce, transportation, communications and energy which for purposes of this paper, have been grouped as services.
However, falling primary activities has not led to the growth of the manufacturing industry, implying that the economy is deindustrializing because construction and services are expanding their respective shares. The standard of living improves and incomes grow because construction and services are activities of "higher quality" than primary ones.
If we determine the speed at which the economy reduces its primary level of activities during the various periods when the phenomenon effectively occurred, we may observe that in the recent economic liberalization period from 1991 to 1997 primary activities fell by 0,25 points per year. In those periods where import-substitution industrialization policies were enforced, the drop was of 1,0 point per year between 1971 and 1975 and 1979 to 1981, and of 1,7 points per year from 1985 to 1987. This means that in the nineties market and liberalization policies lead to lower primarization ratios but at a significantly slower speed than when the process was induced or strategically driven by industrialization policies -in the seventies and eightees.
Such "slow" reduction of primarization during the period when the economy opened up has not been exempt of "macro-economic imbalances" which were also present during the industrialization processes analyzed above. It is worth noticing, for instance, that in 1993 and 1996, the primarization index rose rather than fall (see Figure 16) because of the 1992 drought, in the first case, and due to excess domestic demand in 1994 and 1995 in the second case. This led to macroeconomic adjustment to prevent a further widening of the trade gap. In other words, primarization grows precisely when there is imminent danger of macroeconomic instability and because of prudent short term managing of macroeconomic funda-mentals. If macroeconomic imbalances were not immediately prevented, it is fully possible that primary activities would have diminish further on, thus making possible a strong pendulum come back towards greater primarization at the time the economy finally gets adjusted.
Sustained growth of foreign currency reserves in recent years largely contributed to revert primarization in 1992, 1994, 1995 and 1997. An extremely favorable international environment accounts for larger foreign reserves. Terms of exchange have remained relatively stable, interest rates dropped and capital flows became increasingly global and international. These factors allowed financing a large trade gap. These wider options between 1991 and 1997 permitted to make progress, albeit slowly, towards diminishing the ratio of primary activities to GDP.
It is therefore of particular importance to underscore that macroeconomic stability and good management of macroeconomic fundamentals (whether through Value and quality creation industrialization-promoting policies or in an openly neo-liberal framework), are necessary conditions to achieve the longterm objective of reaching a lower economic primarization index on a sustained basis. Thus improving standards of living.
Peru's case shows though that to reduce the level of primary activities may be much faster if carried out through the active and deliberate design of industrialization policies, as was the case in 1971-1975, 1979-1981, and1985-1987, than if guided by the hand of free markets and liberalization, as occurred in 1991-1997. What went wrong in the last 50 years, was the lack of consistency and coherence between short-term macroeconomic management and industrialization policies. In other words, the type of industrialization policies implemented in Peru failed to prevent trade balance deficits.

ECONOMIC GROWTH
If a country increasingly produces goods with a lower level of processing, quality and value, and on the other hand, consumes foreign goods that include a larger amount of knowledge, undoubtedly it will exchange growing amounts of simple goods for the same (or a smaller) amount of elaborate goods.
If domestic supply, including exports shifts towards primary products with diminishing returns, simple processing or minimum know-how, while demand moves towards increased consumption of sophisticated, complex and always newer products, there will be a gradual uncoupling of quality and value between what the country produces and what it consumes. This uncoupling occurs because not all the economic products and/or activities are alike. It is precisely the more elaborate goods or services that eventually translate into higher standards of living. Elaborate goods incorporate a higher degree of technology and knowledge, show positive externalities, and generate a higher value and have a larger systemic and synergistic effect over the rest of the economy.
Further degradation can be expected conditions will deteriorate even further if the terms of trade fall because of market effects. Declining terms of trade imply prices of exported goods growing more slowly than the prices of imported goods. Although it is true that the terms of trade reflect a difference in the quality and value of economic activities, it is also true that these prices are subject to short term fluctuations that are not related to these factors.
Short term macroeconomic stability and the viability of long term sustained growth become ever more complex if the uncoupling described above is reproduced in supply and demand patterns. This is true to the extent that in this kind of trade increasingly larger exports of simple products will be required to finance similar purchases of more elaborate goods, given the diminishing returns of the first type of products. If under those circumstances exports fail to grow, the emerging external account deficits will eventually hamper all attempts at sustained economic growth, as mentioned in Section 3.2.
Depending on available resources and the country's capacity to produce an increasing amount of simple or primary esan-cuadernos de difusión 92 goods, from a long-term perspective the nation will quite likely end up producing and working more but living under worsening conditions. What evidence is there for Peru? Is the uncoupling of quality and value between supply and demand taking place already? How will this uncoupling hamper economic growth?
A review of the composition of Peru's domestic demand leads to the following remarks: i) Figure 17 shows that imports as a percentage of domestic demand, after having plummeted to 16% at the end of the 1980s, climbed back to 28% in 1997, thus implying that generally speaking Peruvians today consume more imported goods as a percentage of their demand than in the late 1980s.
ii) Table 7 shows that imports of durable and non-durable consumer goods followed by purchases of capital goods for industry recorded the largest relative share increase. These three items together accounted for 46,1% of imports in 1997 while they were only 26,8% in 1988. A difference of 19 percent points iii) On the contrary, the share of inputs and raw materials for industry fell from 43,4% in 1988 to 28,7% in 1997, about 15 percent points less. iv) A flat review of the list of durable and non-durable inputs and of capital goods for industry shows that these goods incorporate relatively more knowledge and require a more complex technical manufacturing process, showing therefore a greater capacity to create value. On the other hand, raw materials for industry are less sophisticated and so create less value.
Some preliminary conclusions may be drawn from the above observations. First Peru is increasing its demand of imported products. Secondly, if we analyze the mix of imports, it becomes obvious that we are importing increasingly sophisticated products with a higher technology or knowledge component (durable consumer goods and capital goods for the industry) while imports of less sophisticated products continue to fall (imports of raw materials for industry). This means that we increasingly buy abroad products of higher quality and value content for which we must pay more.
Such a demand bias is a consequence not only of Peru´s commercial and financial opening but also of increased globalization and communications which is promoting consumption of "cutting edge" or "innovative" products.
At the micro economic level, this demand trend has been confirmed through household surveys that measure increasing ownership of electric appliances in recent years. Table 8 shows not only the growing number of households that own such appliances but also how bottom-end products have been replaced by more elaborate and complex goods. Thus, for instance, color TVs equipped with remote controls have progressively substituted for black and white television sets and those without a remote control. Likewise with the VHS videocassette recording system which replaced BETAMAX or stereo sets that substituted for radios. This is evidence of demand propensities for innovative and improved-quality products that are not produced internally but imported.

esan-cuadernos de difusión 94
These phenomena are not necessarily negative in themselves because progress always comes with increased demand for products that provide a higher level of comfort and quality. The problem emerges when the country has not sufficient capacity to import. At the heart of the matter seems to be the capacity to produce and export goods that are "similar in value and quality" to the imported ones. For this the quality structure of imports should match the value of exports. If this occurs, there is a virtuous circle of growth through international trade that in Graham's model leads to exchange among countries that specialize in activities with constant returns.
But if the structure of production and/ or exports shows that production of pri-mary goods with little or no incorporated knowledge continues to grow, that is of goods with little value added, it is obvious that we will have to work more to meet the observed trend in demand patterns.
A review of the structure of domestic supply (Section 3 in this article) shows that liberalization in the 1990s led to a slower fall in economic primarization when compared to the period of active industrial policy. On the other hand, industrial policies or free market schemes that ignore basic macroeconomic principles, will unfailingly abort if there is lack of coherence with fiscal, external and/or monetary balances.
Empirical evidence shows that when the market is left on its own, economic Source: Apoyo S.A. Prepared by the authors. Value and quality creation primarization diminishes very slowly leading to the emergence of activities not necessarily in the manufacturing sector but in the services (telecommunications, commerce, financial and energy services) or construction industries. In other words, the economy will move towards producing slightly more elaborate non-tradable services due to growing demand by consumers, but never managing to produce the industrial-type tradable goods that consumers demand. Thus, because the demand for manufactured goods cannot be met by domestic supply, those goods are imported. And because domestic production of primary goods exceeds local consumption, they are exported. However, the process originates a growing "uncoupling of quality and value" given that the world's demand for the goods we export grows at a lower pace than our imports. Figure 18 shows that Peru's terms of trade from the beginning of the 1950 to date fell towards the end of 1996 and in 1997 to a fifth of their value at the beginning of the 1970s and currently stand where they were in the mid-1950s.
This means that prices of our exports since 1972 have grown considerably less (or have decreased) compared to the price of our imports, leading to the most substantial fall in standards of living and income ever, excepting the periods from 1975 to 1977, and from 1979 to 1980. In other words, we export products the price of which falls more or rises less than that of our imports, with the subsequent negative effect on domestic revenues and standards of living of Peruvians who have to work more but are worse off.
The question arises then of what made terms of trade improve in the 60s and at beginning of the 70s to then decline since 1972.

esan-cuadernos de difusión 96
An explanation may be found in that at the beginning of the 1960's given productivity levels and demand for materials, inputs and technology, there was an increase in demand and prices of minerals. In the seventies, world oil crisis led to radical technological changes in energy conservation and more efficient use of raw materials which diminished demand and prices of raw materials. Later, in the 1980s and 90s, the communications, electronics and information revolutions introduced massive use of knowledge to produce goods in industrial countries that put Peruvian production and exports at a disadvantage because local industries were unable to incorporate new knowledge at the required speed.
Another hypothesis refers to the extremely poor management of fundamental economic balances since 1970 and the lack of a truly coherent industrial policy that would diversify the supply of exports and promote new activities for which world demand and prices would not fall steadily.
A review of the Japanese experience from 1950 to 1970 shows that the key to higher exports is found not only in their ability to produce better quality exports but principally in their capacity to adjust their export mix by introducing goods for which world demand was growing. In other words, Japan stopped producing and exporting those goods, for which the world's demand was falling and had therefore lower value, and shifted towards production of goods with higher demand.
These are however hypotheses that must be further explored. The concrete fact is that Peru is not producing or exporting the types of good that would ensure and allow it to finance its new consumption structure, a structure that is constantly evolving as the world's demand fluctuates and changes.

Major conclusions
This paper has demonstrated that primarization of economic activities is intimately linked to declining standards of living. For each additional percentage point of primarization, per capita consumption fell by -2,6% while white-collar real wages dropped by -5,4% and bluecollar salaries by -7,4%. The impact on manufacturing industry is in the opposite direction since for each point of increased industrialization, per capita consumption grew by 4,2% and white-collar and bluecollar earnings rose 10,6 and 15,5% respectively (See Section 3).
If wages, salaries and standards of living increased with growing industrialization, why hasn't Peru gradually moved towards the production of manufactured goods? The answer is three-pronged: In the first place, Peru wrongly participates in the world economy basically as an exporter of raw materials and natural resources with diminishing returns.
In the second place, an erroneous industrial policy has favored permanent protection for all types of industries aimed at producing final goods for the domestic market without link ups to other processes and no learning curve. Technology was imported as a package revealing a static business view of the world. There was an insufficient emphasis on education, creation of value and training, few possibilities of tuning up to a broader market, a static rent-seeking attitude, and total despise for "merit-based" promotion. In addition the government sector lacked synergies with private investment and failed to understand the critical role played by the market forces. In summary, an industrial policy illustrated by the slope of the PRO curve in Figure 19.
Thirdly, managing macroeconomic stability and fundamentals is extremely important. Section 3.2 has proven that as the current account, government accounts and/or private deficits became unbearably large, a crisis emerges leading to the corresponding reversal of the industrialization process.
If to these three characteristics we add events on the demand side, a fast-deteriorating situation emerges. Section 4 shows how the structure of demand shifted towards consumption of foreign goods of higher quality, value, knowledge and technology. Morever, if these changes in demand are accompanied by supply that continues to favor the production and export of primary goods, then the country will inevitable end up with a "value or quality mismatch". It will end up working harder to earn the same units of imports. The country therefore specializes in producing and exporting goods with diminishing returns that are exchanged for products with increasing returns. As average productivity falls, so do standards of living (Annex 1 and Sections 1, 3 and 4).
Our most important conclusion is that contrary to most neoclassical theory, we hold that growth and standards of living depend on the types of goods and services produced by a country. A country may master all the exogenous and endogenous factors that are responsible for economic growth, as highlighted by neoclassical theory, i.e. labor, capital, natural resources, savings, investment, technology, efficient use of resources, financial markets, infrastructure, sound macroeconomic indicators, efficient government, etc. However, if all these elements are directed towards producing the wrong types of products, the nation may end up working harder and living worse off.

A framework to design a strategy for growth and increasing standards of living
Peru needs to devote time and energy to "rethink" or review the goods it produces. Figure 19 is a good initial framework for this analysis. The right-hand quadrant shows the reverse postulated relationship between primarization 11 and standards of living. In the left-hand quadrant, we show the positive relationship between the real exchange rate and the primarization ratio, as explained in Figure 18.
Curve NV graphs the negative impact of primary activities on standards of living. It has as implicit parameters a given level of technology, productivity, returns to scale, externalities and other elements that may be grouped as factor "T". Likewise, curve PR describes the positive impact of the real exchange rate on primarization.
It involves parameters related to the type of participation in the world economy and the country's industrial development policy, which depend on the varying focus on the domestic or external markets. These parameters are summarized by factor "C". Thus, for instance, if the country's participation in global trade is based on using natural resources or if the national industrial strategy is basically directed towards its domestic market, PR's slope increases. On the contrary, if industrialization shifts towards promoting manufacturing exports, the slope drops (PR1 in Figure 19).
A close look at these relationships reveals that standards of living would increase, ceteris paribus, by developing an outward industrialization process, which would imply a change in the type of participation in the international economy. In terms of Figure 19, curve PRO would shift downward to PR1 when the strategy to participate in the world economy moves away from the advantages derived from natural resources to the dynamic competitive advantages provided by manufacturing (from C0 to C1). Under these circumstances, given a certain real exchange rate (E0), primarization will fall from PR0 to PR1, from a' to c', and standards of living would increase from N0 to N1 sliding along the NV0 curve, from a to c.
Alternatively, we could introduce a qualitative transformation in primary activities to gradually include technological development in them. Curve NV0 would move to the right, to NV1 (See Figure 20). For the same level of primarization (PO), the population's standard of living would rise from N0 to N1 12 . Technological upgrading (from T0 to T1) in agriculture and mining, for instance, would improve standards of living, given 12 Technological change probably will turn PRO around point a, slightly increasing the impact of the real exchange rate on the degree of primarization. esan-cuadernos de difusión 100 a certain level of primary activities. For this process to be sustainable, technological progress in these primary activities would have to take place on an ongoing basis so as to set off the eventual negative weight of diminishing returns in extractive activities, if this variable is effectively the key factor separating industrial from extractive processes 13 .

REAL EXCHANGE RATE
A third option much in the style of countries with large endowment of natural resources such as Australia, New Zealand and Canada, is to resort to primary comparative advantages at relatively low real exchange rate levels while at the same time promoting manufacturing exports at higher real exchange rates 14 . In terms of Figure  21, this would mean that PR would be an upside down "V", so that at low exchange rates, below EO, primarization would increase. Conversely, with higher rates of exchange, above E0, primarization would fall. The fundamental precept is to curtail total specialization on primary resources and thus prevent the spreading of this sector's diminishing returns throughout the economy 15 .
In this case, by shifting parameter C from CO to C2, starting at point a', an increase of the exchange rate above EO would lead to a declining primarization quotient and to increased industrialization. For this reason, given an E2 exchange rate, if manufacturing output has been efficiently directed toward either the external or domestic market, manufactured products would be competitive and primary products would be significantly profitable. However, the higher profitability of the primary sector would have to be offset to prevent new entrants or expansion of incumbents in areas of diminishing returns. Usually, countries with abundant natural resources that have successfully increased their population's living standards have regulated and restricted the use of their resources for both strategic and environmental reasons. In this case, developing a complementary manufacturing sector, with either an inward or outward orientation depending on the size and volume of the market, is crucial for any such economy to improve its standards of living.
Obviously, it is possible to combine the three approaches mentioned above or try other paths which will modify the implicit parameters on both the left and right quadrants.
One initial proposal for Peru would make it devote itself to primary activities at low real exchange rate levels preventing the country from entering into areas with diminishing returns when the real exchange rate increases. As real exchange rate rises, it should speed up promotion of a strong export-oriented manufacturing sector very much in the style of the inverted "V" curve shown in Figure 21. Rents should be extracted from primary activities when real exchange rate increases like in Canada, Australia and New Zealand, as a way to limit their expansion 13 We must bear in mind that a key element separating primary from manufacturing activities is the former's diminishing returns and the latter's increasing returns.
14 It is also possible to develop a manufacturing sector on the bases of the internal market if this is sufficiently large to benefit from economies of scale as Friederich List proposed in the last century. Value and quality creation and address environmental concerns. The other two approaches appearing in Figures  19 and 20 could also be followed but with much more limited results.

Free markets and strategic supply policies in a global world
How should Peru therefore face the new era of globalization and free trade that is promoted by multilateral organizations and the industrial economies?
Assuming that "free markets" by themselves will automatically lead Peru to the kind of specialized productive structure implicit in point d' of Figure 21 above is a fallacy that was tested in the liberalization of the 1950s and 1990s. On the other hand, assuming that any kind of industrial policy will ensure that this path will be followed is another fallacy tested in the 60s, 70s and 80s. Peru needs to design productive strategies that will consciously and deliberately allow us to "create" and "ensure" permanent increasing returns and dynamic competitive advantages over time, by producing higher value goods and services and maintaining well-balanced macroeconomic fundamentals.
Peru must therefore enter the new global and trade era not necessarily following the dictum of multilateral organizations and developed economies but ensuring the production of those goods and services that will generate faster growth and increased standards of living.
Neither free trade and free markets per se, nor industrial and commercial policies that are inconsistent with market forces will prove successful in this venture. To improve Peru's standards of living there is a need to combine strategic policies and market orientation in order to create dynamic competitive advantages in the production of high value/quality goods. In an initial situation of autarchy (case 1), Country A produces and consumes 800 units of wheat and 800 units of watches. Average productivity of both activities is 4 units per worker-day and the labor used in each activity is 200 worker-days.
Country B produces and consumes 800 units of wheat and 600 units of watches. Wheat productivity is 4 units per worker-day and that of watches is 3 units per worker-day. Each activity employs 200 worker-days.
Based on these assumptions, the world's wheat production would be 1.600 units and 1.400 watches or a total production of 3.200 units in wheat terms, given the relative price of the two products. Country A's product is 54% of the world's production (or 1.714 wheat units) while Country B accounts for 46% of the total (or 1.486 wheat units).
When opening trade (case 2), and given their respective comparative advantages, Country A specializes partially in watch making while Country B turns to wheat production. Presumably, Country A will transfer 100 worker-days from wheat cropping to watch making, while country B transfers labor in the opposite direction, i.e. from watches to wheat.
Since both activities show constant returns, world's watch output should increase by 100 units while wheat's remains stable. Consequently, global world production in wheat terms grows by 114 units, or 3,6%.

Case 1: Without trade
Increased world production (equivalent to the added balance of commerce between the two countries) is distributed equally, thus demonstrating that the specialization through trade, given each nation's comparative advantages, is beneficial for both countries, to the extent both activities show constant returns.
Case 2: With trade + Specialization by comparative advantage.
Country A transfers workers from wheat to watches. Country B transfers workers from watches to wheat. + Constant returns in both activities.
However, if wheat has diminishing returns and watches, increasing returns (case 3), international trade, based on comparative advantages, will be beneficial only for Country A, i.e. product will grow by 16,3% when compared to the autarchic situation, while B will hurt from a 13,9% production fall.
In this example, the world's wheat production drops by 100 units, and that of watches grows by 150 units. Consequently, product and world trade, in wheat terms, will grow in 2,2% (or 71 units).
Graham's model predicts that the country specializing in production of goods with diminishing returns will experience both, a drop in its GDP and a trade gap. Just the opposite will happen to the country specializing in goods with increasing returns. That is, if one of the activities shows either increasing or diminishing returns, trade will hurt one of the two countries. Otherwise said, trade will not be equal if there are differences in returns.

Should wheat have diminishing returns and watches constant returns (case 4), trade specialization reduces Country B's GDP leading to a trade gap. To wit, trade will adversely impact the nation with comparative advantages in the good with diminishing returns.
The world's production would grow very slightly, and so would commercial exchanges, and, in general, the gain in Country A almost equals Country B's loss. Country a transfers workers from wheat to watches. Country B transfers workers from watches to wheat. + Constant returns in watches and diminishing returns wheat.
If watch production shows increasing returns and that of wheat, constant returns (case 5), the final solution would be quite similar to the previous example, since the production in Country B would fall and in Country A it would grow. A's trade balance would show a surplus while B's would be negative.
In spite of this similarity, the world's product and commercial trade would grow more than in the previous case.