The Secret is in the Brew: No Quantitative Approach to Institutional Performance Can Go it Alone

Cynthia Taft Morris, Smith College

I. Introduction

The thesis of this paper is a common sense one: No single methodological approach can explain economy-wide economic performance. I focus on institutional changes determining performance because economic change only happens through institutions. Hence, economic performance is institutional performance. Institutions provide economic agents with the incentives and institutional channels necessary in complex societies for productive economic activities. The literature on how well institutions perform - whether in the aggregate or in capital or labor markets - takes little account of performance in distributing benefits from economic growth. This is in spite the fact that most investigators, when pressed, agree that raising per capita GDP should raise the standards of living of the majority of the population. Quantitative studies of economy-wide economic performance typically fail to include noneconomic variables. The reasons, seldom given, are probably that noneconomic variables are assumed less important than economic ones or are too closely correlated with economic ones for inclusion or cannot be quantitatively measured.

The purpose of this paper is to demonstrate that explaining economy-wide economic performance requires a minimum of three complementary methods of analysis: quantitative time-period analysis; cross-section analysis of countries sharing broadly similar resource configurations, which includes at least some key noneconomic variables; and historical study of relevant path-dependent feedback interactions especially those involving economic growth rates and political institutions.

To illustrate these points, I shall discuss five land-rich settler economies before World War I, starting with a selective review of time-series studies of individual countries. I then move to the Morris and Adelman cross-country analysis of the late 19th and early 20th centuries, Comparative Patterns of Economic Development, 1850-1914 (1988) to illustrate the complexity of associations between per capita GDP growth and immigration, on the one hand, and land and political institutions, literacy, government policies, and foreign dependence, on the other. I shall finish by illustrating how the past history of cumulative domestic political-economic feedbacks involving per capita GDP growth and immigration helped determine contrasting patterns of national economic performance and immigration. (I discuss feedbacks involving foreign dependence in a companion paper.)

II. Time-series studies of per capita GDP, capital, and immigration patterns.

This section will review selected studies of individual settler economies before World War I, focusing on explanations of patterns of per capita GDP and capital and immigration flows. The section underlines illustrates my view that these studies underestimate the importance of influences shown in cross-section and comparative historical studies to influence levels of per capita GDP, immigration, and the distribution of benefits from economic growth.

III. A cross-section study highlighting the role of institutions in growth, immigration, and distribution.

This section will illustrate why time-series studies of individual countries fail to explain adequately either economy-wide economic performance or changing levels of per capita GDP and immigration. The Morris and Adelman cross-section results for five settler economies before World War I highlights contrasting institutional features of successful economic growth, on the one hand, and successful widespread distribution of growth's benefits, on the other. Political influences, land distribution, and the extent of primary education for example, weighed more heavily on distributional outcomes than on economic growth rates.

IV. Why time-series and cross-section studies together are not enough: the path-dependent approach.

This section is an interpretation of some of my findings in a qualitative historical study comparing the five settler economies referred to above (published in the 1992 volume of Research in Economic History.) In the early phase of capitalist growth in these economies, strong path- dependent political-economic feedbacks cumulatively altered institutional constraints on labor systems in ways that were fundamentally different in Latin American and British settler economies. These feedbacks contributed to contrasting mixes of labor systems with different distributional characteristics and contrasting patterns of government intervention in land, labor, and credit markets. They interacted with regularities in economic and demographic responses to market forces and "chance" events to cause persistent contrasts between the Latin American and the British settler economies in per capita GDP, levels of living, and distributional patterns.

The path-dependent approach is best known for applications where chance events determine selections among competing techniques subject to increasing returns that are unbounded relative to the size of the market of potential adopters. With costs that decrease through cumulative learning effects or network or coordination externalities, the technology selected first by chance gains a cumulative cost advantage over alternatives, which can prevent later adoption of a potentially superior technology and "lock- in" an inferior technology first selected. Similarly, chance initial location of a city can generate such strong cumulative dynamic externalities that it prevents later choice of a potentially superior site.

The application of the path-dependent approach to sets of complex interdependent political, economic and institutional changes is not straight-forward. In the countries studied here, alternative political scenarios rarely involve mutually exclusive economic institutions or techniques. Rather feedbacks between macro-level political and micro-level economic influences cumulatively change the mix of economic institutions by altering the costliness of alternative institutional arrangements. In the settler economies studied here, political decisions determine constraints to which politically influential economic agents respond economically at the micro level. At the same time they act politically at regional or national levels to alter the constraints in ways benefiting them economically at the micro level.

For simplicity, I assume that potential institutional changes in the settler economies studied here can be roughly delineated by the full five sets of actual subsequent changes during the remaining decades before World War I. Political-economic feedbacks changed the probabilities of micro- level "selections" from these alternative institutional arrangements in three ways: (1) they raised greatly the transactions costs of adopting subsets of "possible" institutional innovations or rearrangements which politically dominant economic groups found not to benefit them; (2) political-economic feedbacks increased the efficiency characteristics of politically preferred institutional adoptions through the benefits of learning-by-using, coordination and other dynamic externalities, and political actions fostering both; and (3) they increased the interdependence of predominant choices of technique, predominant choice of labor system, and predominant choice of type of labor in ways raising product shares of politically dominant economic groups.

Thus, viewed ex ante and across countries, individual enterprises faced politically determined constraints on "potential" institutional changes that could and did vary greatly among settler economies. These politically determined constraints contributed to differential rates of expansion by type of enterprise and holding, which in turn led to contrasting aggregate mixes of enterprise and holding types, each having different distributional characteristics, with consequent systematic variations in aggregate income distributions.

In the remainder of this section, I lay out two contrasting stylized models of political-economic feedbacks that constrained institutional change and thereby circumscribed the distributional consequences of economic growth. These feedbacks are one element in an overall explanatory scheme that must include the traditional well-understood regularities summarized by economic modeling as well as (1) the chances of geography, soils, climate, and location; (2) the chances of discovery and initial settlement, initial distribution of land and other assets, and initial sociopolitical structures; and (3) the chances of external events, politics, and individual actions.

V. Stylized models of political-economic feedbacks.

I propose two contrasting historical stories about the role of self- reinforcing political-economic feedbacks in constraining micro-level institutional selections. These stories offer a framework for understanding key features of contrasting distributional patterns, which can encompass considerable diversity of country experience.

The Latin-American story. Typically, in Latin America a dominant pattern of large estates was established long before the period studied here. Powerful large landholders had long dominated local political structures. As Western Europe industrialized and population increased sharply, expanding European purchases of overseas food and raw materials, helped greatly by new ocean steamships and refrigeration, attracted European investment in crops, marketing, railroads and financial institutions. Influential Latin American intellectual elites adopted European visions of a liberal society based on independent family holdings. They promoted the homestead programs introduced to attract Western European immigrants, which offered credit, abundant land, and help with passages. Parliamentary systems were widely introduced but the lack of secret ballot and fraudulent electoral procedures helped large estate owners retain their political dominance, particularly at regional and local levels. So rapid was the increase in European demand for primary exports that large estates could not obtain sufficient labor from the families of homesteaders and found or perceived alternative sources of domestic labor to be insufficient. Governments stopped or reduced homestead programs and in line with the shift in the preferences of large estates shifted their advertising to low-income immigrants, usually from Southern Europe, who often sought quick gains for remigration home rather than permanent settlement.

The consequences of the shift from homestead to low-income labor were economic and political. Estate owners responded to rapidly expanding market opportunities with innovative cost-reducing tenant and wage-labor systems and technical improvements that increased their share of overall export production. Private national and regional credit markets developed for mortgages for large estates and for short-term commercial loans to local merchants. Local credit became available to tenants from local merchants for commercial short-term loans at high interest rates, which supplemented advances from estate owners. Mortgage markets for homesteads failed to develop. This lack of private credit for homestead acquisition combined with the cessation of public land and credit for new homesteads greatly raised the transactions costs of acquiring new homesteads and led to a falling domestic market share of homesteads in overall export production. These rising transactions costs also reduced the alternatives to wage and tenant work, increasing both the monopsony power and product shares of large estate owners and the profits of local merchants. Immigration of low-income workers much greater than in British settler economies held wage levels down.

The shift to low-income estate labor lacking both education and political awareness and a marked disinclination of parliaments to provide effective electoral procedures led to very low levels of immigrant citizenship and political participation. The cessation of public homestead programs, rising transactions costs for new homesteads, and falling homestead domestic market shares combined with the lack of effective channels of political participation to limit the political influence of homesteaders.

Thus in this stylized model of Latin-American political-economic feedbacks, initial land concentration and political dominance were reinforced by political success in obtaining low-income tenant and wage- labor for large estates and in limiting public land, credit, and infrastructure in the homestead sector during the rapid economic expansions of the late 19th and early 20th centuries. The reinforcing through political mechanisms of both economic and local and regional political dominance of large estates thus helps account for the persistence of comparative inequality in the distribution of assets and income during the economic booms before World War I.

While these cumulative feedbacks operated to reinforce institutions that generated inequality of assets and income, The process of economic expansion gradually increased the middle class and demands for political participation, thus ultimately weakening the feedbacks and eventually altered their nature after World War I.

The British-settler story. The second story is one in which large landholders initially dominating regional parliaments with wide suffrage and the secret ballot lost political power to two groups. Enfranchised British immigrant workers, in the aftermath of the gold rushes that had attracted them, pressed for immigration restrictions and wider access to land; while merchant and manufacturing groups in response to domestic economic growth pressed for regional public investments benefiting them. Varying coalitions of merchants and manufacturers, smaller-scale farmers, and workers became politically dominant, pressing for: (a) immigration restrictions against low-income countries; (b) universal public primary education; (c) public investment, wage, and social policies in effect favoring unskilled workers; and by the end of the century (d) public land, credit, and agricultural policies that favored family-scale farms. The small and middle-size agricultural enterprises so favored responded to international competition with innovations that used both capital and land more intensely. Heavy reliance on British immigration resulted in a degree of cultural homogeneity that helped reinforce restrictive immigration policies, which in turn strengthened the degree of political consensus about preserving incomes and benefits of rising economic groups, which jointly (or in turn) dominated the political process and as individuals benefited in local labor and land markets from political influence in regional or national parliaments. (This story will be expanded in the final paper.)

VI. Some preliminary conclusions.

The feedbacks outlined above appear to reconcile the apparent conflict between within-country studies in which micro-level institutional changes appear neatly neoclassical and cross-country comparisons in which they do not.

Changes in financial, marketing, labor, and land institutions were broadly neoclassical in the intensely competitive international environment of the decades before World War I. For individual countries, aggregate immigration and capital flows varied along lines predicted by standard economic theories of international trade, migration, and investment. But cross-country studies show contrasting patterns of change in per capita GDP, immigration, and foreign dependence as well in contrasting levels of per capita GDP, agricultural productivity, and access to land, credit, and education. In addition, the terms on which alternative institutional selections were available (determined by the contrasting political-economic feedbacks described in the paper) varied greatly across countries and help explain the persistence over time of contrasting levels of per capita GDP and contrasting proportions of the population benefiting from economic growth.