LUMPERS AND SPLITTERS IN ECONOMICS, A NOTE.

Kindleberger, Charles P.

Charles P. Kindleberger [*]

In refereeing a paper about financial crises which used the Minsky-Kindleberger model, I encountered the remark that on this subject I am a "lumper," presumably one who views all financial crises as more or less alike. A few days later in idly reading an obituary of someone I did not know, Sir Eric Stokes, a historian of British India, I learned that Sir Eric's demonstration that preexisting Indian political systems and British agrarian taxation had created very different farming outcomes in different parts of north India, was very much the work of a "splitter" rather than a "lumper." (Proceedings of the British Academy, 1997). An essayist, Ann Fadiman in Ex Libris (1998) writes that "George [her husband] is a lumper. I am a splitter." The vocabulary was new to me. The Shorter Oxford English Dictionary (SOED) provides appropriate definitions of each word and relates them, one to the other.

"A lumper is a person (esp. a taxonomist) who attaches importance to similarities rather than differences in classification or analysis, and so favors inclusive categories (Cf SPLITTER);" while a splitter is "a person (esp. a taxonomist) who attaches importance to differences rather than similarities in classification or analysis, and so favors subdivision (Cf. LUMPER)."

The dating system of the SOED indicates that "lumper" came into use in 1830-1869, whereas "splitter" was later: 1870-79.

I gather that the term lumper is not pejorative, and grant that it is more or less true that I believe financial crises have broad similarities. In the contemporary East Asian case, the crises in Thailand, Indonesia, Malaysia, and South Korea were caused by financial deregulation, herd behavior among Asian borrowers and European and North American lenders, fixed exchange rates as expanding quantities of money produced inflation, and with the rising U.S. dollar against the yen, overvalued currencies. It is a puzzle, however, how to weigh the similarities and differences. The December 1998 issue of the Asean Economic Bulletin, for example, has a paper by Andrew McIntyre on domestic political institutions in Thailand and Indonesia which are different though the crises proved to be similar. Thailand's government was honest, weak, indecisive; Indonesia's corrupt with strong and decisive command.

Lumper and splitter, or lumping and splitting are words not common in the social science I read, but seem useful. A new book on Global Transformations (1999) divides analysts on the subject among "hyperglobalizers," "sceptics," and "transformationalists." Hyperglobalists think the organization of the world is entering a new era of political and economic integration and cultural convergence and that the nation-state will atrophy. Sceptics believe that the nation-state will survive and that economic interdependence goes back a long way. Transformationalists take the view that the matter is more complex and that most predictions will be wide of the mark, with separate functions changing in separate ways. The first two categories seem to be lumpers; the last splitters. In this discussion, the authors, David Held and three other British, appear to imply that splitters are intellectually superior to "hyper-globalists" and "sceptics." I am reminded of a statement of a late M.I.T. physicist who claimed that "Everyth ing is more complicated than most people think."

In international relations at the moment, both lumping and splitting are taking place in political organization. There is lumping as the European Union, NAFTA, Mercosur and Asean indicate, and splitting, unhappily in Yugoslavia, and potentially in Quebec, the Basque country, Scotland, Wales. The process of making big ones out of little ones, and little ones out of big ones seems endless. Scholarship deals with separate countries, and larger agglomerations such as the Mediterranean, Latin and Central America, Africa, north and south of the Sahara desert. In economic history William Parker and Sidney Pollard believe in treating the coal vein stretching from the Pas de Calais in France, through Belgium to the Ruhr as an economic integer. Regionalism--a salient topic today--can be thought of as splitting if the counterfactual with which it is compared is globalism, or lumping if the alternative is the nation-state.

Economics has often been torn by the question whether it is better to aggregate or disaggregate. Lumping and splitting are not an exact parallel, since some issues involve the contrast between two lumps, or between a proper lump and an aggregation which should be split, or even one between two split groups. There is in economics a series of bipolar issues which illustrate these possibilities: public choice and governmental regulation against free markets, Keynesianism v. monetarism, centralization of authority v. pluralism, rules v. men who make decisions on the basis of particular circumstances, consumer and property-owner freedom of action but constraints on behavior with possible negative effect on others such as traffic lights, zoning, and some regulations, often ignored, such as seat-belts for those riding in automobiles.

The philosopher, Karl Popper, believed in parsimony, explaining outcomes with the smallest possible number of causes. This is a form of lumping. Opposed is such a physicist-philosopher, as Ilya Prigogene, who votes for complexity, often a series of causes, none of which is sufficient, but all, or most, necessary. The Santa Fe Institute, headed by W. Brian Arthur, is home to the notion of complexity. And the economist, Albert O. Hirschman, produced a paper a few years ago with the title "Against Parsimony."

Hirschman also has a book, Exit, Voice and Loyalty (1970), exit standing for the economic response to something disliked or not wanted--quitting the job or not buying the product; voice, speaking up within an organization without leaving when one disagrees with the consensus of the community or the decisions made. Exit and voice are both lumps, reactions that are highly similar. Loyalty, on the other hand, modifies each. It means not exiting, but staying on the job or continuing to buy the product despite dissatisfaction; it also modifies voice by keeping silent even in disapproval. Originally exit was split from voice. Lately, in Crossing Boundaries (1998), Hirschman has backed away from this separation, suggesting that the surge of the crowd's voice in East Germany in 1989 led to widespread movement abroad into Hungary, and then to the West, a spectacular form of exit which ended in the collapse of the Berlin Wall and the unification of East and West Germany.

Lumping and splitting are sometimes completely separate in economics, sometimes interrelated in complex ways, as in the East German case. Intermediation, for example, between buyer and seller in entrepot markets not connected with either, or between creditor and debtor both depending on banks for lack of adequate information, has moved on to disintermediation, firms trading directly with one another without the help of lumpy centers. Sweden will buy wool in Australia or Spain rather than in Antwerp, Amsterdam or London, and industrial companies in need of finance will sell certificates of deposit or even whole bond issues directly to insurance companies and other cash-rich firms rather than to banks or the capital market.

The choice between lumps and splits may vary with the function involved, and may change as circumstances evolve. In the usual historic case, banks were started locally but gradually formed into financial centers through mergers or change of physical location, centers which produced a hierarchical ordering covering wide areas as Lyons, Paris, Amsterdam, London, New York, and increasingly Frankfurt, illustrate. Insurance companies also developed pyramidal structures, sometimes separate from banking as in such cities as Hartford and Munich. Other agglomerations occurred initially in retail distribution (before the widespread use of the automobile encouraged shopping malls with large areas for parking) and in theatre districts--but not motion pictures apart from Times Square and Picadilly, and in financial dealings, like Wall Street and the City. Changing circumstances dampened some of this movement: a number of New York banks moved up from Wall Street to mid-town to nestle with the head offices of large nationa l and multinational corporations, an initial lumping followed by a split. A similar split seems to be in process in banking in the rise of Raleigh, North Carolina into a major financial center, some distance from largescale industrial or other financial activity. I hypothesize that this development is owed to the cheapening of communication which has cut down to a degree on the need for face-to-face business dealing.

There may or may not be need for change from lumps to splits as scale increases. On the side of "not," English agriculture was dominated by the nobility throughout the process of enclosures and "territorial amalgamation" from 1780 to the 1820's, until especially the Reform Act of 1832. In Aspects of the Aristocracy (1994), David Cannadine explains that English nobility--itself expanded by "peers, courtiers, statesmen, nabobs, royal physicians and naval and military commanders"--plus a number of self-made men, bought estates in these years in an active market for land, popularly known as "Terramania." The price of land rose from 23 years' purchase to 28 years; rentals in England by 70 to 90 percent, these owing partly to the high price of wheat. When that price started to decline after the Napoleonic wars, it was propped up by passage of the Corn Laws.

Ownership of land by nobles also led to lumping in mineral exploitation, especially of coal, and, through a linkage, to canal building. The much higher scale of finance needed for railroads from the later 1830's induced splitting. M.C. Reed's Investment in Railways (1975) notes that capital was contributed from a number of unrelated sources: land owners, often noble, whose territory was needed in part for rights of way; industrialists such as Boulton and Watt, or Josiah Wedgwood, who hoped to gain from cheaper and safer transport for inputs and output; investors of all sorts seeking income; speculators who observed the rise in prices of railway equities; vendors of railway equipment who would occasionally accept partial payment in railway shares; and by an intermediating route, the government. As an example of this last, John Gladstone, father of the later prime minister, William, was paid [pounds]93,526 in 1835 by the British government as compensation for the freeing of slaves in his plantations in Demerar a and Jamaica. With his son, he put [pounds]70,000 into shares of the Grand Junction Railway, and some money into the Forth and Clyde canal to bring that investment to [pounds]40,000 (S.G. Checkland, The Gladstone Family, 1971). In these early days, only one bank, Glyn Mills, participated in lending to railways (Leland H. Jenks, The Migration of British Capital to 1875, 1927).

Technical change and rising scale increased the need for capital, and altered its provision in lumps from connected persons. In early modern times, ships were financed by the captain, his family, friends and neighbors, often all from the same port. With the move from small sailing to large steam vessels, that system became obsolete and the catchment basin had to be extended. The need for capital by manufacturers at the start of the industrial revolution, Sidney Pollard has asserted, was small. Manufacturers used rented buildings, credit provided by suppliers, and family and neighborhood money to pay workers. As scale and technology changed, however, new means were required. The process can be illustrated in cotton textiles. The sharply enlarged demand after the cotton famine of the U.S. Civil War led to the increased size of the average cotton mill, which grew from 10,500 spindles in 1862, to 15,600 in 1871, and in Oldham in the 1880's to 65,300 spindles. Mills were owned by joint-stock companies, raising ca pital by selling shares in stock exchanges rather than within the family, i.e. splitting rather than lumping. (J.B. Jeffreys, Trends in Business Organization in Great Britain since 1856, L.S.E. dissertation, 1938; published 1977).

If one looks at economic theory rather than historical practice, one finds both lumping and splitting. In early discussion of the current account in the balance of payments, a scholar such as Fritz Machlup leaned heavily on price elasticities, a lumpy theory, while Sidney Alexander stressed "absorption,"--how much of national income was spent on consumption and investment and how much saved, the difference between savings and investment representing a deficit in the current account if investment exceeded savings, and a surplus if vice-versa. Harry Johnson produced a third, and not very plausible explanation, comparing the demand for money and its supply at the national level: net demand created a favorable balance as the country exchanged goods and services for additions to money supply, or spend the excess money on import of goods and services if money was redundant. While each explanation was a lump, the existence of three resulted in an overall split. Today's record U.S. current-account deficit is ascribe d to the complex interaction of a wider list of factors: low personal savings (despite the reduction of the government deficit), falling world prices, the flight to quality of foreign holders of wealth seeking the safety of dollars, recession or depression in Europe and Asia. As far as I am aware, no econometrician has tried to put weights on these separate forces within the overall split.

Economic historians who search the past for patterns of behavior are on the whole lumpers, of whom the most notable, perhaps, was Karl Marx. There are, however, a long list of lumpy classes of economic behavior in given areas, perhaps the Marxian generalization, refined by Sir Arthur Lewis as "growth with unlimited supplies of labor," Engel's law that with rising income the proportion of income spent for food and other basic necessities declines, the law of one price that in a single market for a single good, there will be but one price, Gresham's law that bad money drives out good.

Economics also has lumpy pairs of ideas, within a relative narrow class, in conflict one with the other. As noted earlier, I have in mind public v. private goods, Keynesianism v. monetarism, centralization v. pluralism, rules and regulations v. men, exit v. voice, goods that are properly marketed v. those that philosophers insist should not be traded: people, honors, political office, freedom of expression, criminal justice, freedom from military service and jury duty, and more (Michael Walzer, Spheres of Justice, 1983). In many historic instances, and some contemporary, these "goods" have been bought and sold against money. There are other goods and services that are found on both sides of these split categories: public and private education, for example, or public and private hospitals and health care. In a form of cognitive dissonance, holding two opposing views simultaneously, such as a strong belief in private property but favoring restrictions on owners of real estate, say, in a historic district, to a lter their structures without oversight authorization, is a particular kind of split.

Perhaps the most far-reaching gap between lumps and splits in economics lies between the micro- and macro- branches of the subject. Macro-economics has within it the split between Keynesianism and monetarism, and within the former that between changes in public and private spending, the latter as affected by tax changes. J. Kenneth Galbraith, for example, wanted expansionary fiscal policy under the Kennedy administration effected through public works rather than tax deduction, presumably because invested capital produces a stream of output over time whereas lower taxes may merely increase consumption of evanescent objects. But macro-economics is on the whole more lumpy than micro-economics. There may be choices among measures to affect the net national product, the price level and the exchange rate, and between general measures and those targeted at particular functions such as bank regulation or liberalization. There may also be splits because of the index-number problem, and in the difference between nomin al and real income, the latter not always carefully observed. Micro-economics, however, seems to me to fall into many more split categories.

Micro deals with markets, let me say, and there is the Arrow-Debreu equilibrium which, provided the assumptions necessary to its functioning competition, profit-maximizing producers and intelligent and informed buyers, are met, is self-equilibrating. The assumptions, however, are on the whole unrealistic. Markets may be subject to monopoly, duopoly or oligopoly. As between producers and consumers, information may be asymmetric, and procrastination on the two sides of the market may be unbalanced. Trade may be subjected to high transport costs, high transactions costs, cognitive dissonance, positive or negative externalities. There is also the possibility of market failure--though financial crisis probably belongs under macro-economics, with a need for a lender of last resort which poses the danger of moral hazard. Karl Polanyi and Robert Solow insist that the work of labor should not be commoditized.

Markets normally perform magnificently, as illustrated by the complex planning that went into providing for the needs of the armies invading Normandy in June 1944 contrasted with the effortless supply of New York's needs daily by supply responding to demand.

My instinct tells me, however, that an economic taxonomist would find more reasons for splitting micro-economics than macro. I might perhaps mention that Harry Johnson reviewed James Meade's The Balance of Payments (1951) which won its author the Nobel Prize in Economic Science in 1977, calling it taxonomy, as if this was a criticism. (In due course, Johnson became a great admirer of Professor Meade and organized a conference in his honor). Taxonomy is defined by the SOED as:

"Classification, esp. in relation to the laws or principles of the branch of science, or a particular science or subject that deals with classification, esp. the classification of living organisms.

It is perhaps doubtful that economics or other social sciences deal with living organisms; the terms taxonomy or classification do not appear frequently in these fields. But the difference between a lumper and a splitter in social science in general may be worth pursuing.

I have frequently called attention to the split between an "optimal economic area" (enlarged from Robert Mundell's optimum currency area) which is, in my judgment, the world, and the optimal social area which is far smaller. The criterion of economic optimality is the equalization of prices for goods and services and factors of production (for similar classes); an optimal social area is one small enough to give its members a sense of belonging to a cohesive identity. It is hard to see how economics and sociology, for example, can be lumped as optimal for a given area.

The only sensible reaction to this insight, new to me, but not I judge to historians, is to adopt both lumpers and splitters as useful. Taxonomy (classification) is efficient as an early step in social science, and the notion of more than one brand is helpful.

(*.) Professor of Economics, Emeritus, M.I.T