Specialization and Regulation: The Rise of “Professionals” and the Emergence of Occupational Licensing Regulation

2nd Draft—Please Do Not Quote September 22nd, 2003

Marc T. Law* and Sukkoo Kim**

Abstract Why did university-educated professionals come to dominate so many occupations in late nineteenth and early twentieth century America? And why was the rise of professionals often accompanied by licensing regulation? We argue in this paper that the expansion of specialized knowledge contributed to the growing importance of university-trained professionals, and that similar forces also led to the adoption of occupational licensing laws. ______________________________________________________________________ * Assistant Professor, Department of Economics, University of Vermont, Old Mill Building, 94 University Place, Burlington, VT 05405-0114, E-mail: Marc.Law@uvm.edu ** Associate Professor, Department of Economics, Washington University, Campus Box 1208, One Brookings Drive, St. Louis, MO 63130-4899, E-mail: soks@wueconc.wustl.edu

Not long ago, the Governor of a midwestern state was approached by representatives of a particular trade anxious to enlist the Governor’s support in securing passage of legislation to license their trade. “Governor,” the men said, “passage of this licensing act will ensure that only qualified people will practice this occupation; it will eliminate charlatans, incompetents or frauds; and it will thereby protect the safety and welfare of the people of this state.” The governor, from long experience, was somewhat skeptical. “Gentlemen,” he asked, “are you concerned with advancing the health, safety and welfare of the people under the police powers of the state, or are you primarily interested in creating a monopoly situation to eliminate competition and raise prices?” The spokesman for the occupational group smiled and said, “Governor, we’re interested in a little of each.” (Council of State Governments, 1952, p. 1)

I. Introduction

The term “profession” seems to lack a clear definition. According to the Merriam-

Webster Dictionary (1997, p. 585) a profession is “a calling requiring specialized

knowledge and often long academic training.” Taking this definition a step further,

William Rothstein (1972, p. 8) argues that a profession is “a manner of earning a

livelihood through the application of a body of highly abstract knowledge in some set of

institutions.” An even more detailed definition of the word “profession” has been offered

by Paul Starr, who writes (1982, p. 15): “A profession… is an occupation that regulates

itself through systematic, required training and collegial discipline; that has a base in

technical, specialized knowledge; and that has a service rather than a profit orientation

enshrined in its code of ethics.”

While each of these definitions differs in terms of its emphasis and level of detail,

a common theme emerges: all agree that professions are occupations that sell services for

which specialized knowledge is an important input, and that extensive training may be

needed to acquire this specialized knowledge. From an economist’s perspective, the

emphasis on specialized knowledge is important because it suggests that the services

offered by “professionals” may be characterized by asymmetric information about

product quality. If professionals sell services that require specialized knowledge (and

extensive training), then presumably it is not an altogether straightforward task for

ordinary consumers to judge the quality of these services. Hence, to this list of definitions

we add our own: “A profession is an occupation that sells specialized services whose

2

product quality is difficult to verify ex ante by the buyer.” Product quality is difficult to

verify ex ante precisely because specialized knowledge is a critical input in the

production of professional services.

Throughout history, uncertainty about product quality has obviously been a

common feature of many goods and services. Nevertheless, it was not until the late

nineteenth and early twentieth centuries that modern day professionals—individuals who

acquired a university education as an integral part of their vocational training—came to

dominate so many occupations. During this period, formal education at universities and

colleges gradually displaced apprenticeships with experienced practitioners as the

primary means of vocational education for a wide range of specialized occupations.

Additionally, the emergence of these professionals was accompanied by extensive

regulation. Why did university-trained individuals become so dominant in occupations

where specialized knowledge was important? And why was the rise of these

professionals accompanied by state regulation?

We argue in this paper that the expansion of specialized knowledge during the

nineteenth and twentieth centuries made it advantageous for individuals engaged in many

professions to acquire a systematic, formal education. As the size and scope of the

knowledge base that was required to function effectively in a variety of occupations

expanded, ad hoc learning through apprenticeship under an experienced practitioner

could no longer provide adequate training. Hence, universities and colleges—which

offered systematized, formal, instruction in a variety of disciplines—gradually became

the dominant route by which individuals acquired the skills and knowledge to work in

occupations for which specialized knowledge was an important input.

We also argue that the expansion of specialized knowledge, combined with

urbanization and the rise of impersonal market exchange, contributed to the emergence of

occupational licensing laws. As markets became increasingly anonymous, and as

specialized knowledge became a more important component of professional services,

asymmetric information about quality became even more severe. This situation generated

a potentially productive role for occupational licensing regulation as a solution to the so-

called “lemons” problem that can arise when sellers have better information than buyers

about product quality (Akerlof 1970; Leland 1979).

3

Existing scholarship on occupational licensing regulation generally emphasizes

the role that licensing plays as an entry barrier (Moore 1964; Friedman 1962; Kessel

1970, 1972, Maurizi 1974). As is well known, entry barriers are valuable to incumbent

producers because they reduce the extent of competition, allowing incumbents to increase

their prices and profits (Stigler 1971). What is less well acknowledged, however, is that

entry restrictions can also play a role in reducing informational asymmetries. As the

quotation at the beginning of this paper suggests, producers may find licensing laws

productive as a devise that helps solve the “lemons” problem and serves monopolistic

purposes. We believe this to be a fundamental dilemma in the regulation of professional

occupations.

This paper is structured as follows. In section II we examine how the expansion of

specialized knowledge gave rise to the dominance of university-trained professionals in a

variety of occupations and how changes in the nature of the market for professional

services increased the severity of the asymmetric information problem. In section III we

provide information on the emergence of occupational licensing laws in the late

nineteenth and early twentieth centuries. In section IV we attempt to test the two

hypotheses by examining the effect of licensing on entry into a variety of occupations

and by conducting an event history analysis of the adoption of licensing legislation for

the same set of occupations. Since data limitations make it difficult to conduct a decisive

econometric test, case study evidence of the history of a particular occupation may be

helpful. Hence, in section V we examine the history of the medical profession and the

emergence of medical licensing laws. Finally, section VI concludes the paper.

II. Expansion of specialized knowledge and the rise of university educated professionals

It is clear that scientific knowledge expanded tremendously during the nineteenth

and twentieth centuries. The scope and depth of human understanding about the physical,

social, and biological world underwent a massive expansion during these centuries. Not

only did the total stock of knowledge grow, but scientific knowledge also became

increasingly specialized. This was reflected within academic institutions, as well as

within the scientific literature, which became increasingly focused around specific

disciplined. “Science” as a general subject area was gradually replaced by “physics,”

4

“chemistry,” and “biology” and these particular disciplines were slowly subdivided into

even more specific sub-disciplines (Price 1963).

The expansion of specialized knowledge can be illustrated by examining the

growth and changing composition of the scientific literature. Table 1 presents data on the

number of scientific periodicals published in a variety of scientific disciplines between

1780 and 1900. The data shown in this table are taken from Gascoigne (1985), who

surveyed and categorized by scientific field approximately 900 scientific periodicals

published in Germany, Great Britain, Italy, and the United States throughout the

eighteenth and nineteenth centuries. The data reveal a dramatic increase in the number of

scientific periodicals published over time, an increase in the number of field-specific

scientific journals, and a decline in the share of “general science” journals. The total

number of scientific journals in Gascoigne’s sample increased from 30 in 1780 to nearly

800 in 1900. However, the portion of those journals that were classified as “general

science” periodicals declined sharply from 90 percent in 1780 to 27.5 percent in 1900.

Much of the new scientific knowledge that was acquired over this time was published in

specific field journals. Disciplines like botany, geology and physics that only had a

handful of specific periodicals in 1800 had several by the turn of the century. Indeed,

according to Gascoigne’s sample some disciplines, like zoology, had no periodicals until

1820 but had over 100 by 1900. Thus, there appears to have been an expansion in both

the scope and depth of scientific knowledge during the nineteenth century.

These trends continued throughout the twentieth century. The top panel of Table

2 displays estimates of the number of periodicals within six scientific disciplines:

astronomy, biology, chemistry, geography, mathematics, and physics. To estimate the

number of periodicals published within each discipline we consulted Fowler (1966), who

compiled a list of periodical guides for a variety of academic and professional fields as

well as information on the number of periodicals included within each guide. For each

discipline we examined, the number of specialized journals exceeded 600 by the mid-

1900s. Hence, during the twentieth century, the accumulation of specialized knowledge

and the expansion in the total stock of knowledge continued apace.

Over time, new scientific knowledge found applications in occupations like

medicine and engineering. Advances in chemistry, physics, and geology during the

5

nineteenth century made it possible to develop specialized fields within engineering.

Specialized engineering fields like chemical, mechanical, electrical, geological and

metallurgical engineering gradually came into being. A systematic understanding of

increasingly specific scientific disciplines was integral to the development of these new

engineering fields and formed an essential part of the training of new engineers (Society

for the Promotion of Engineering Education 1930). Meanwhile, in medicine, the

development of basic biological sciences like bacteriology, microbiology, immunology,

and biochemistry made it possible for physicians to identify the sources of particular

disease. While significant improvement in the ability of physicians to treat many of these

diseases only emerged gradually, a systematic understanding of biological science

became important for the training of physicians, veterinarians, and other health workers

(Flexner 1910; Ludmerer 1985).

As scientific knowledge became useful to practitioners, a scientific literature

devoted to applications in medicine, dentistry, engineering, and architecture and directed

to practitioners of these occupations also began to emerge. Hence, the expansion of

specialized knowledge occurred not merely within “pure” science fields, but also within

occupations where scientific advances and the scientific mode of inquiry proved useful.

Over time, specialist journals directed at practitioners within these occupations

proliferated. The lower panel of Table 2 shows estimates of the number of scientific

periodicals published within five “scientific occupations.” By the middle of the twentieth

century, there were approximately 2,700 “current” periodicals in medicine, 550 in

engineering, and 570 in veterinary medicine. Thus, the twentieth century also witnessed

an advance in specialized knowledge that was particular to certain occupations.

The consequences of this explosion of specialized knowledge were twofold. First,

it became increasingly difficult to be a successful generalist. While it may have been

possible in the 1500s for an exceptional individual like Leonardo da Vinci to be an

engineer, architect, and physician, it was clearly not feasible to be a master of more than

one of these fields by the early 1900s. Therefore, within particular academic disciplines,

as well as within particular occupations, specialists emerged who possessed a deeper

knowledge of a more specific subject area or application. Second, the benefits of longer

and longer periods of formal education increased the pre-requisite knowledge base

6

needed to gain a mastery of certain fields increased. Completion of primary or even

secondary schooling and mastery of “reading, writing, and arithmetic” ceased to furnish

an adequate educational background for increasingly specialized and scientific

occupations (Society for the Promotion of Engineering Education 1930; Ludmerer 1985).

Universities, colleges, and other institutions of higher learning therefore emerged to

educate individuals who wanted to acquire the knowledge and skills to work in

occupations where the growth of relevant knowledge was greatest.

Table 3 presents data that document the expansion of higher education in the

United States from 1870 to 1970. Over this period there was a substantial increase in the

number of institutions of higher education, the number of students enrolled in these

institutions, enrollment as a percentage of the 18-23 year old population, and the number

of degrees granted from these institutions. While there were undoubtedly many factors

that were driving the growth of higher education, part of this upward trend must be

attributable to the comparative advantage possessed by university educated professionals

in an environment where knowledge was increasingly specialized.

The effects of specialization on the market for professional services

Economists are quick to note that specialization increases the gains from trade.

By allowing each of us to become “more different,” specialization raises the benefits of

exchange. However, when each of us is more different, transaction costs must be

overcome if we are to capture the benefits from trade (Wallis and North 1986). Perhaps

the most important of these transaction costs are those related to the acquisition of

information about product quality. Specialization increases information costs because as

each person becomes more specialized, the less she knows about the nature of the goods

and services she purchases from others. The information set each individual possesses

about the goods purchased from others shrinks, partly because the goods themselves may

become more complicated, but also because specialization may result in the introduction

of new and more familiar products. Specialization may therefore create an asymmetric

information problem in which buyers know less about product quality than sellers and

this situation can give rise to the so-called “lemons” problem where low quality goods

dominate the market (Akerlof 1970).

7

The expansion of specialized knowledge and the emergence of university trained

professionals during the late 1800s and early 1900s were accompanied by significant

changes in the nature of the markets for the services offered by these professionals. In the

late eighteenth and early nineteenth centuries, the bulk of the population lived in rural

areas or in relatively small towns. While many of the services offered by physicians,

dentists, and other professionals were subject to uncertainty, in traditional small town

markets local reputations could play an important role in solving the asymmetric

information problem. According to Starr (1982) and Rothstein (1972), markets for

physicians’ services during the colonial and early antebellum periods were relatively

“thin.” There were generally very few physicians in each town, and high transportation

costs made it impractical to go beyond one’s local community in search of medical care.

Hence, simple reputation mechanisms could work reasonably well in helping individuals

judge the quality of physicians.

As the nineteenth century progressed, however, falling transportation costs

brought about the integration of the national economy, which in turn contributed to

regional specialization and urbanization (Kim 1995, 1998, 2000). This movement of the

population out of the countryside and into increasingly dense cities was accompanied by

the rise of impersonal exchange as the dominant form of market interaction.

Specialization and the rise of impersonal exchange created problems for producers in a

wide variety of markets. In the food industry, for instance, urbanization and the rise of

impersonal markets for foodstuffs contributed to growing uncertainty about food quality,

since consumers knew less and less about how their food was produced and what was

added to it (Law 2003). For sellers of professional services, the rise of impersonal market

exchange posed similar problems since local reputations became less effective as

exchange became increasingly anonymous. Membership in key professional societies or

associations may have been a partial substitute for local reputation, but in an environment

where professional societies were proliferating in nearly every occupation and the

requirements for membership in these societies were often very low the signaling value

associated with membership in any given society may not have been particularly high.

Hence, specialization and the growth of impersonal markets throughout the nineteenth

8

century made it increasingly difficult for professionals to establish a reputation for good

service.

The expansion of specialized knowledge and its application in a variety of

professions also served to exacerbate the problem of asymmetric information. In

medicine, for instance, the development of tools like the microscope, ophthalmoscope,

and laryngoscope dramatically increased the ability of physicians to diagnose diseases,

but how were patients unschooled in scientific medicine to know if they had received the

correct diagnosis? Each of these tools made it possible for the physician to “see things”

that ordinary individuals could not; and without the ability to independently verify and

understand what physicians were doing, it was harder for the patient to evaluate the

quality of the diagnosis. Hence, by increasing the complexity of many professional

services, the growth of specialized knowledge also made it harder for consumers to know

what they were getting.

Thus, by the late nineteenth century, professionals in many occupations found the

markets for their services dramatically altered. Urbanization and the rise of impersonal

markets made it more costly for professionals to establish reputations for high quality

service. In addition, the expansion of specialized knowledge increased the complexity of

many services, which in turn made it more difficult for consumers to know what they

were getting. Hence, the changing nature of the market for many services required

professionals to search for new ways to solve the asymmetric information problem.

III. The emergence of occupational licensing regulation in the late nineteenth century

Occupational licensing in America did not begin in the late 1800s. During the

colonial period and into the early nineteenth century, state governments regulated the

practice of law and medicine in various ways. Nevertheless, it seems reasonable to assert

that occupational licensing on a wide scale did not begin until the late nineteenth century.

During the latter half of the 1800s and continuing through early 1900s, state governments

throughout American enacted legislation that regulated a wide variety of occupations.

While the substance of these regulations varied from occupation to occupation and from

state to state, in general these laws set standards determining what it would involve to

become a qualified practitioner of a particular occupation. In some instances, these laws

9

also made it illegal for those who did not meet these standards to practice particular

occupations. This second wave of licensing regulation differed from colonial licensing

regulation in several ways. First, it applied to a wider variety of occupations than

medicine and law; the licensing statutes enacted in the late 1800s and early 1900s

included occupations ranging from barbers and beauticians to architects, engineers, and

dentists. Second, the requirements that were established by these regulations were

generally stricter than those that were in place during the colonial era. Although colonial

era licensing laws generally did not preclude unlicensed professionals from practicing

medicine or law, in the late nineteenth and early twentieth century a license was

generally a pre-requisite to practice a given occupation. Third, the enforcement of these

more recent regulations was probably more vigilant than in the colonial era (Council of

State Governments 1952).

Table 4 presents information on the timing of licensing laws for a large sample of

occupations. While state governments began to regulate some occupations as early as the

1870s, most occupational licensing laws were adopted in the first two decades of the

1900s. Physicians and dentists were among the first occupations to be licensed by state

governments during this second wave of state licensing regulation. Licensing laws

regulation physicians and dentists were first adopted in the 1870s. By the early 1900s,

most states had enacted some kind of medical or dental licensing regulation. In contrast,

the earliest laws regulating accountants, architects, and nurses were generally not enacted

until the 1890s and early 1900s. For these occupations, it was not until the 1910s and

1920s that the majority states had enacted licensing legislation. Hence, the timing of

licensing varied considerably across different occupations. Finally, while licensing laws

were enacted for some relatively low skill occupations (barbers and beauticians, for

instance) who sold relatively simple services, most of the occupations that became

regulated were “professions” in the sense that they sold specialized services whose

product quality was difficult to verify.

Not surprisingly, the story behind how licensing laws came to be adopted varies

from occupation to occupation, as well as from state to state. Nevertheless, the historical

evidence suggests a few plausible generalizations (Council of State Governments 1952).

One is that licensing was generally preceded by the formation of local, state, and federal

10

professional societies and associations. From the mid-1800s onward, professional

societies proliferated in fields like architecture, civil engineering, dentistry, medicine,

and pharmacy. While the motives behind the emergence of these societies were

numerous, their primary objectives were to maintain standards of professional conduct

and training, to providing a forum for the dissemination of new practices and techniques,

and to prevent the entry of “unqualified” practitioners. In medicine, for example, medical

sects with different therapeutic approaches formed their own medical societies at local,

regional and federal levels. Associations representing the medical “mainstream” were the

most numerous, but during the mid-1800s, as the popularity of “heroic” therapy (i.e.

bloodletting, etc.) declined, there was a proliferation of medical societies representing

physicians who advocated different therapeutic approaches. Each of these societies had

their codes of conduct, promoted different therapeutic philosophies, and desired to

control entry into their profession in various ways (Rothstein 1972). In mechanical

engineering, how practitioners were trained formed the basis for different mechanical

engineering societies. Mechanical engineers who were trained in mechanist shops formed

their own societies. Those who were trained at engineering colleges belonged to different

societies. While each of these societies played an important role in the development and

diffusion of new engineering techniques, they also disagreed over the type of training

was needed to become a qualified engineer (Calvert 1967).

Another important generalization is that the main political impetus for licensing

generally came from some subset of professional associations within a given occupation.

The leaders of particular professional associations and societies were often among the

most vocal proponents of licensing regulation. Within medicine, for instance, political

pressure for licensing regulation initially generally came from certain state and local

medical societies who represented the medical mainstream (Rothstein 1972). A third

generalization is that, within each occupation, not all practitioners or professional

societies were in favor of licensing regulation or of the particular form of licensing that

was being advocated. While some subset of profession was in favor of regulation, other

segments were indifferent to regulation or even hostile to it. In law, attorneys who

practiced in the largest New York firms were apparently indifferent to proposals to

tighten bar requirements (Stevens 1983). In medicine, disagreement among different

11

medical sects about the validity of particular therapies divided physicians. These

divisions often led to conflicts about the desirability of medical licensing laws (Starr

1982; Rothstein 1972).

Because different professional societies often disagreed about either the

desirability of licensing or the particulars of proposed licensing laws, the licensing laws

that were enacted often represented a compromise among the different segments of the

profession. Grandfather clauses exempting existing practitioners from new educational

requirements were frequently included in order to obtain the support of older members of

the profession. When competing segments of profession felt that licensing laws would be

used against them, efforts were made to include these groups in the organizations charged

with enforcing the licensing statute. This was almost invariably the case in medicine. In

many states, boards that included representatives of each of the major medical sects

enforced the licensing laws that were enacted; alternatively, separate licensing boards

were established to regulate each medical sect. According to Rothstein (1972, p. 309),

“cooperation among the sects was almost always absolutely necessary to obtain licensing

legislation in states with many non-regular physicians.” Hence, the licensing laws that

were ultimately adopted generally reflected the views of a relatively broad cross-section

of the occupational group being regulated.

IV. Why did licensing arise? An empirical analysis

Did state licensing arise to restrict entry into professional occupations? Or did it

emerge to help solve an asymmetric information problem? In this section, we attempt to

bring some systematic evidence to bear on this issue. Specifically, by estimating the

effect of licensing laws on the number of individuals working in different occupations, as

well as by examining some of the factors that influenced the enactment of state licensing

regulation, we hope to shed some light on this debate.

An examination of the effect of licensing laws on the number of individuals

working in a variety of occupations furnishes us with a reasonably direct test of the entry

barrier hypothesis. If licensing laws were an entry barrier that reduced the extent of

competition in the market for certain services, then other things equal, the per capita

number of individuals working in a particular occupation should decline, or the growth

12

rate in the number of individuals in a particular occupation should fall following the

adoption of licensing legislation.

An event history analysis of the determinants of licensing regulation may also be

informative because it provides some rough indication of the motivation for state

licensing. If the adoption of licensing legislation for a given occupation was only

influenced by membership in that occupation, then it would appear that licensing was

introduced primarily to advance producer interests. Producer interests may have desired

licensing either because it served as an entry barrier, or because licensing helped solve

the asymmetric information problem. A positive correlation between producer interests

and the licensing is therefore consistent with both the entry barrier hypothesis and the

asymmetric information hypothesis since professionals may desire licensing either to

deter entry or to solve the lemons problem (or both). On the other hand, if the enactment

of licensing legislation is also correlated with proxies for the severity of the asymmetric

information problem facing a particular occupation, the evidence is more strongly

suggestive of the possibility that licensing was introduced to reduce informational

asymmetries. A positive correlation between the extent of asymmetric information and

licensing legislation is therefore consistent with the asymmetric information hypothesis

but not with the pure entry restriction hypothesis.

Effect of licensing regulation on entry

To examine the effect of licensing laws on entry, we gathered state level data

from the Census of Population on the number of workers in several different occupations

for each census from 1870 to 1930 and we matched this data with information about the

timing of state occupational licensing laws. This data set allows us to exploit cross-state

variation in the timing of occupational licensing laws to estimate the effect of regulation

on the per capita membership in various occupations and the growth rate of membership

in these occupations. We restricted our attention to the 1870 to 1930 censuses because

the occupational classifications over these years were most consistent. After 1930, the

Census Bureau significantly altered the definitions of various occupations, which limits

the comparability of the post-1930 data with the pre-1930 data (Edwards 1942). This is

13

unfortunate, but not fatal for our purposes since, as shown in Table 4, by 1930 most

occupations were already regulated by most states.

We were able to find state-level data on the number of persons working in each of

the following eleven occupations: architects, attorneys, barbers, beauticians, dentists,

engineers, nurses, physicians, plumbers, teachers, and veterinarians. Data on the timing

of state occupational licensing laws that regulated each of these occupations was taken

from the Council of State Governments (1952). Since our data have a panel structure, it is

possible for us to use a fixed-effect framework for our empirical analysis. In particular, if

we let Pit denote the population of state i in year t; let Lit denote the number of workers in

a particular occupation in state i in year t; let Si denote a fixed-effect for state i; let Yt

denote a fixed-effect for year t; let Rit denote a binary variable that is equal to 1 in all

years t in which occupational licensing is in place in state i and 0 otherwise; and let Xit

denote a vector of control variables for state i in year t (for instance, income per capita in

state i in year t, or the urbanization rate in state i in year t), we can determine the effect of

occupational licensing on entry by estimating, for each occupation, ordinary least squares

fixed-effect regressions of the following form:

(1) Lit/Pit = µ + Si + Yt + αRit + Xitβ + εit

The variable of interest in this regression is α, the coefficient on the regulation dummy

variable. If α < 0 then regulation reduced the number of individuals employed in a

particular occupation, which suggests that regulation functioned as an effective entry

barrier. If, on the other hand, α ≥ 0, then regulation either had no impact on entry or

increased the entry of individuals into a particular occupation.

We can also examine the effect of occupational licensing on entry by estimating a

“dynamic” version of equation (1), which takes the following form:

(2) (%∆Lit) = γ + Si + Yt + δ(%∆Pit) + αRit + νit

In regression equation (2), the dependent variable is %∆Lit, which is the growth rate

between adjacent census years (approximated, for our purposes, by the first difference in

the natural logarithm) in the number of individuals working in a given occupation. As

before, α is the coefficient of interest. If α > 0, then occupational licensing regulation

reduced the growth rate of the number of persons in a given occupation (i.e. reduced the

14

rate of entry), which implies that the licensing regulation served as an entry restriction.

On the other hand, if α ≥ 0, then occupational licensing had no impact or increased the

growth rate of the number of persons working in a given occupation (i.e. did not affect or

increased entry), which indicates that the occupational licensing law was not an effective

entry barrier.

Table 5 presents the ordinary least squares regression estimates of equation (1).

For each occupation we present two sets of estimates. The first controls for the

urbanization rate in each state. The second controls for both the urbanization rate in each

state as well as the real per capita personal income in each state. Since personal per capita

income data at the state-level are only available for 1880, 1900, 1920 and 1930, for the

second set of regressions we estimated the equations limiting the sample to these

particular years.

For most occupations, the coefficient on the regulation dummy variable is not

statistically significant, although the sign and magnitude of this variable does vary across

the occupations. For beauticians and teachers, the coefficient on the regulation dummy

variable is positive and statistically significant, suggesting that regulation increased entry

into these occupations. For physicians, architects, and dentists this coefficient is negative

but not significant. For engineers, attorneys, and veterinarians, the regulation dummy is

positive but not significantly different from zero. The control variables (the percent of the

population that is urban, and income per capita) in each of the regression equations

generally have plausible signs. For nearly every occupation, urbanization is correlated

with a higher number of workers per 1,000. Higher personal income per capita is also

correlated with more workers per 1,000.

An analysis of how state licensing laws affected the growth rate of the number of

workers in each occupation may be a more useful way to estimate the impact of licensing

on entry, since our intuitive notion of entry may correspond better with changes in

growth rates than changes in levels. Hence, we now turn our attention to Table 6, which

reports the fixed-effect regression estimates of equation (2) for each occupation. In this

set of regressions, the coefficient on the regulation dummy variable is negative but not

statistically significant for architects, barbers, dentists, engineers, and plumbers; positive

but not significant for teachers, attorneys and nurses; positive and significant for

15

beauticians; and negative and significant for physicians and veterinarians. Thus, while

licensing may have functioned as a weak entry restriction for physicians and

veterinarians, it did not restrict entry into the other occupations examined.

Event history analysis of the adoption of licensing laws

To determine what motivated state governments to enact licensing regulation, we

conduct an event history analysis. In particular, we estimate hazard regressions of the

adoption of licensing legislation for several occupations. Hazard models are appropriate

for this purpose because they allow us to exploit the longitudinal structure of our data set.

Within the class of hazard models, we use the logistic discrete time hazard model. This is

a sensible choice because the adoption of licensing legislation was a single, non-repeated

event (at least within our data set), and because our observations are separated by ten-

year intervals along the temporal dimension (Allison 1984).

The regression equations we estimate take the following form:

(3) log[Rit/(1 – Rit)] = λ+ θ(Uit) + π(Lit/Pit) + ρ(Lit/Pit)2 + εit

Rit is binary variable that is equal to 1 in the year t in which occupational licensing for a

particular occupation is adopted in state i and 0 otherwise. Lit/Pit is the per capita number

of individuals working in a given occupation in state i in year t. Uit is the urbanization

rate in state i in year t; and εit is an error term. The per capita population of individuals

working in a given occupation is included to proxy for producer interests. As noted

earlier, producers may desire licensing because it functions as an entry barrier and/or

because licensing may help solve the asymmetric information problem. We also include

the square of the per capita population of workers in a given occupation to allow for non-

linearity in the effect of occupational group size. While increases in the per capita size of

the occupational group may lower the time until licensing is adopted (i.e. increase the

licensure hazard), the effect of group size may eventually become negative because free

riding or coordination problems among members of the group reduces its political

effectiveness. The urbanization rate is included as a very rough proxy of the degree to

which exchange was impersonal in each state, and hence, the extent to which asymmetric

information was problematic. If urbanization raises the licensure hazard, then we have

evidence that is consistent with the asymmetric information hypothesis, but not with the

16

pure entry barrier hypothesis. For each occupation, we estimated the regression equation

with and without year-specific binary variables that allow the hazard rate to vary over

time.

Table 7 displays the coefficient estimates from this exercise. While the coefficient

estimates do vary somewhat, depending on whether the hazard rate is allowed to vary

from year to year, the regressions indicate that urbanization had a positive and

statistically significant effect on the licensure hazard for architects, attorneys, physicians,

and veterinarians. For the other occupations, the coefficient on urbanization is generally

positive but not statistically significant. For barbers, beauticians, dentists, engineers,

nurses, physicians, and plumbers, the per capita number of individuals in each occupation

has a positive and statistically significant effect on the licensure hazard. In general, the

coefficient on the square of the per capita number of individuals in each occupation is

negative, implying that the effect of occupational group size on the licensure hazard

diminishes as the size of the group expands beyond some threshold.

These coefficient estimates should be interpreted with caution, since the

urbanization rate and the per capita number of individuals in a given occupation are very

crude proxies for the underlying forces they are intended to represent. Keeping this

caveat in mind, the regression results suggest that for nearly every occupation, producer

interests were an important factor influencing the timing of licensure, and that for certain

professions (architecture, law, medicine, and veterinary medicine), the extent to which

exchange was anonymous, approximated by the urbanization rate, also affected the

timing of licensure. Since historical evidence suggests that producer groups were

generally the main proponents of licensing, the statistical significance of producer

interests in most of the hazard regressions is not altogether surprising: producer interests

probably stood to gain the most from licensing, either because they perceived it would

function as an entry barrier or because of its potential to reduce informational

asymmetries and help make a market for certain service. Hence, both hypotheses posit

that licensure should be a function of producer interests. That urbanization also had a

positive and significant effect on the licensure hazard, however, provides stronger

evidence for the asymmetric information hypothesis than the entry-barrier hypothesis, at

least in the context of these particular professions. Hence, our event history analysis

17

suggests that, at least for certain professions, licensing laws were adopted to solve the

asymmetric information problem that arose as markets became increasingly impersonal.

The skeptical observer may still be unconvinced since, as noted earlier,

urbanization is a very poor proxy of the extent to which the asymmetric information

problem was present. Indeed, the same skeptic might also note that since urbanization

may be correlated with political activity more generally, a positive correlation between

urbanization and the licensure hazard provides very little information about the validity

of the asymmetric information hypothesis. While we are sympathetic to the first criticism

and wish that we had better data that would allow us to conduct a cleaner test, we believe

that two factors militate against the second criticism. The first is that the urbanization rate

did not have a positive and statistically significant effect on the licensure hazard for all

occupations. The second is that the coefficient on the urbanization rate is only positive

and significant for relatively “sophisticated” professions. Indeed, it is noteworthy that for

barbers and beauticians (two relatively low-skill occupations for which asymmetric

information about quality may not have been very important), the coefficient on

urbanization is not significant. If urbanization also had a positive and statistically

significant effect on the licensure hazard for barbers and beauticians, the skeptic’s case

would be stronger because asymmetric information was unlikely to be problematic for

these occupations. Since it is not, however, we believe that the positive and significant

effect that we do find in other regressions furnishes some information about the validity

of the asymmetric information hypothesis.

V. A case study of the medical profession

Given that data limitations make it difficult to conduct a clean econometric test

that allows us to distinguish among the competing hypotheses that have been offered to

explain the emergence of licensing regulation, it makes sense to turn to case study

evidence. Hence, we now provide a close analysis of the history of the medical

profession and of medical licensing laws. We focus our attention on physicians since a

large body of information is available about the history of this profession, and because

physicians are often singled out as the paradigmatic example of a licensed profession.

18

Medical practice and medical licensing in antebellum America

During the colonial era, the best physicians in America were trained in either

Britain or continental Europe. The small handful of physicians who were domestically

trained generally obtained their education by apprenticing with experienced physicians.

Although the nature of medical practice throughout the colonial period and for much of

the nineteenth century was relatively primitive, asymmetric information about physician

quality was still a concern for consumers and producers of medical services. Indeed,

higher quality physicians from the colonial period onward recognized the need to solve

the problem of asymmetric information and attempted to organize the market for their

services in ways that allowed them to distinguish themselves from lower quality

competitors. While in rural areas the relatively thin market for physicians’ services made

it possible for skilled doctors to develop a local reputation as the “county physician,” in

urban areas where the number of doctors was much higher, physicians trained in Europe

or who apprenticed under the best physicians generally belonged to local medical

societies that set the standards for professional conduct and training. Membership in

these medical societies was often restricted to those physicians who met certain

educational requirements or who were considered among the elite of the profession in a

given city. Examples of such societies include the Boston Medical Society, which was

founded in 1780 to create “a just discrimination… between [physicians] who are duly

educated, and properly qualified for the duties of their profession, and those may

ignorantly and wickedly administer medicine whereby the health and lives of many

valuable individuals may be endangered, or perhaps lost to the community” (quoted in

Rothstein 1972, pp. 65-66) and the College of Physicians in Philadelphia which,

according to Shyrock (1960, p. 152) was established in 1787 to set some physicians

“apart from other practitioners in the community.” Exclusivity was a key feature of both

the Boston Medical Society and the College of Physicians: the charter of the Boston

Medical Society initially restricted the number of members to seventy, while the College

of Physicians set prohibitively high membership fees and charged large annual dues.

Indeed, the exclusivity of these organizations is what made membership in these societies

valuable, since only the “best” physicians could belong to such elite medical societies.

Hence, membership in an elite medical society conveyed information about physician

19

quality, and joining such societies was the mechanism by which the best physicians

distinguished themselves from the rank and file of the profession.

State governments enacted medical licensing legislation during the late 1700s and

early 1800s, generally as part of legislation that either incorporated or established state or

county medical societies (Rothstein 1972; Kett 1968). These statutes varied in form, but

in general gave medical societies the right to issue licenses and to determine the

prerequisites needed to obtain a license. While these early licensing laws may have

played some role in solving the asymmetric information problem (since to qualify for a

license, new physicians usually had to complete a three year apprentice with an

experienced practitioner), they did little to block entry into the medical profession. In

many states, a license merely conferred the physician with a right to sue patients for non-

payment. In these jurisdictions, nothing prevented unlicensed physicians from practicing

medicine. In those remaining states that outlawed the practice of medicine without a

license, the penalties that could be imposed on unlicensed practitioners were generally

low. Indeed, the available historical evidence suggests that in these states, unlicensed

practitioners were seldom punished for violating the state’s medical licensing regulations.

Hence, Rothstein (1972, p.76) writes that “[w]hile legislatures were generally willing to

grant licensing powers to medical societies, they were unwilling to enact laws which

would have seriously deterred unlicensed practitioners.”

As the nineteenth century progressed, the profession’s interest in licensing

declined and, in the 1830s and 1840s, nearly every state repealed its medical licensing

legislation (Rothstein 1972; Starr 1982; Ludmerer 1985). Two factors contributed to the

decline of licensing during this period. First, during the early to mid-1800s, the

acceptability of traditional medicine, with its reliance on bloodletting, massive doses of

dangerous metals, and other forms of “heroic” therapy, waned as alternative and less

aggressive therapeutic approaches (i.e. Thomsonianism, Eclecticism, and Homeopathy)

became popular with physicians and patients. With traditional medicine discredited, the

value of having a license issued by a society that advocated traditional medicine also fell.

Hence, membership in established medical societies found their numbers depleted as

physicians joined new medical societies and medical sects that promoted alternative

therapies.

20

Second, the early 1800s witnessed the proliferation of proprietary medical schools

that offered an alternative to apprenticeship as the basis of medical training. While many

physicians continued to be trained by apprenticing under an experienced practitioner, the

quality of apprenticeships varied considerably since not all physicians had the time,

resources, or inclination to offer their pupils systematic training in basic scientific

subjects like anatomy, physiology, and biology. To fill this void, physicians began to

open proprietary medical schools that offered more systematized instruction in key

medical subjects (Rothstein 1972; Kett 1968). As the popularity of proprietary schools

increased, the physicians who operated these schools were able to persuade licensing

authorities to automatically grant licenses to medical school graduates. While the

proprietary schools did remedy some of the deficiencies of the apprenticeship system,

their physician-owners generally had little incentive to offer high quality training since in

most jurisdictions the possession of an M.D. degree from a proprietary school

automatically conferred a license to practice medicine. It thus became a profit

maximizing strategy for the owners of proprietary medical schools to churn out as many

graduates as possible at the lowest cost, which resulted in the quality of medical training

at most schools becoming notoriously low. In general, the pre-requisites for attending a

proprietary medical school were low or even non-existent, the course of instruction was

seldom longer than two-four month terms (most apprenticeships, in contrast, lasted for at

least three years), written examinations were infrequent, and virtually no clinical

instruction was offered (Ludmerer 1985). Growth in the number of proprietary medical

schools and proprietary medical school graduates during the early 1800s thus undermined

the value of medical licenses, since graduates of even the worst medical schools were

entitled to a license.

The re-emergence of medical licensing in the late 1800s and early 1900s

Medical licensing laws re-emerged in the late nineteenth and early twentieth

centuries, largely (but not exclusively) at the prompting of the profession itself. What

sparked this renewal of interest in state licensing legislation among the medical

profession? We believe that two main factors drove the profession to lobby for licensing

legislation in the late 1800s and early 1900s. The first was the rise of “scientific”

21

medicine. The second was the emergence of the modern medical school. We now turn to

an investigation of how these factors contributed to this second wave of medical

licensing laws.

The rise of bacteriology, biochemistry, immunology and microbiology as

scientific disciplines in the latter half of the 1800s and the growing acceptance of the so-

called “germ theory” of disease had profound implications for the nature of medical

practice and the viability of licensing legislation. First articulated by Pasteur in the 1860s,

the germ theory led to the discovery of the microbial origins of a number of illnesses

including cholera, diphtheria, pneumonia, tetanus, and dysentery (Ludmerer 1985).

Although the implications of the germ theory were not immediately obvious to most

physicians, its usefulness eventually became apparent to physicians from all the major

medical sects. Hence, the rise of “scientific” medicine, by providing a common ground

for physicians trained in “traditional,” “homeopathic”, or “eclectic” approaches, reduced

the extent of sectional conflict over the appropriate treatment for particular diseases, and

helped reunite the medical profession. State, local and national medical societies that

used to exclude members of alternative medical sects gradually became more inclusive,

and the political effectiveness of these organizations rose correspondingly (Rothstein

1972, Starr 1982). This in turn facilitated collective action on the part of the medical

profession that made it possible for physicians to obtain licensing laws that set some

standards for entry into the profession.

The rise of “scientific” medicine also helped forge a consensus as to how

physicians ought to be trained. By demonstrating the usefulness of the methodology and

knowledge base of the biological sciences for medical practice, the growing acceptance

of scientific medicine forged a permanent link between academic science and medicine.

This in turn contributed to changing the profession’s attitude toward the nature of

medical education (Ludmerer 1985). Abraham Flexner, a Progressive Era education

reformer who wrote an influential report on the state of medical education, echoed the

attitudes that emerged from this new consensus. In this report, Flexner (1910) argued that

scientific medicine required physicians to have a firm grasp of the basic sciences, and to

be able to apply the scientific method to medical practice. He emphasized that the

training of scientific physicians required modern medical schools with well appointed

22

laboratories, academically qualified students, and faculty that were actively engaged in

medical research. Hence, a university-based medical curriculum slowly emerged in the

late nineteenth and early twentieth centuries in order to equip medical students with a

thorough understanding of bacteriology, biochemistry, immunology, pathology, and other

biomedical sciences.

The high cost of providing this “scientific” medical education made it unlikely

that proprietary medical schools would initiate improvements in the quality of medical

training since higher costs implied that tuition fees in proprietary schools would have to

increase significantly. Most of the major improvements in medical education during this

period appear to have taken place in non-profit medical schools (for instance, Johns

Hopkins) that were generally affiliated with major research universities (Ludmerer 1985).

These medical colleges were able to take the lead in raising the quality of medical

education, partly because universities were able to underwrite the costs of their facilities,

but perhaps more importantly, because being a non-profit, university-affiliated institution

served as a signal of quality to potential donors and to prospective medical students.

Improvements in medical education thus took place in these institutions rather than their

proprietary counterparts since their non-profit, university-affiliated status signaled to the

marketplace that they were more concerned about producing high quality physicians than

about making a profit from operating a medical school.

In spite of these developments, proprietary medical colleges continued to thrive

throughout the late 1800s, and the evidence suggests that these colleges persisted in

offering low quality medical training. Between 1860 and 1880, the total number of

medical colleges in the United States expanded from 65 to 100; in the next two decades,

the number of medical colleges increased even more rapidly, reaching a total of 165 by

1900. This increase was also reflected in the number of medical school graduates, which

rose from 3,241 in 1880 (the earliest year for which we have data) to 5,241 in 1900

(Council on Medical Education 1920). While it is not possible for us at this stage to

determine what portion of this expansion was due to the proprietary schools, the available

historical evidence suggests that the continued expansion of proprietary schools was

largely to blame. Although it may have been apparent to leaders of the profession which

medical colleges trained good doctors and which did not, in an increasingly urban an

23

impersonal society, this may have been far less obvious to ordinary consumers. Thus, the

licensing laws that were enacted during this period took aim at the lower end of the

medical education market, and gradually introduced standards that raised the quality of

medical education and forced the lowest quality proprietary schools to shut down.

According to data from the Council on Medical Education (1920), the number of medical

colleges in America declined from 160 to 131 between 1900 and 1910. By 1920, this

number had fallen to 85. Hence, even before the Flexner report of 1910, licensing laws

began to play an important role in improving the quality of medical education by forcing

proprietary schools to either raise their standards or close their doors (Starr 1982;

Ludmerer 1985).

Although our empirical analysis indicates that state medical licensing laws during

the late 1800s and early 1900s operated as a relatively weak entry restriction, other

scholarship suggests that in subsequent decades, medical licensing laws functioned more

successfully as an entry barrier. Friedman and Kuznets (1945), for instance, show that

physicians’ incomes rose dramatically during the 1920s and they attribute this increase to

medical licensing laws that strengthened organized medicine’s ability to control entry

into the medical profession. While this is undoubtedly true, we believe that the pure entry

restriction hypothesis misses an important part of the story. The medical licensing laws

that were enacted by state governments during this era clearly increased the monopoly

power of physicians over time, but by creating barriers to entry in the medical education

market that forced proprietary medical schools to close down, they also helped solve the

lemons problem. Physicians presumably sought licensing legislation because of its

potential to serve both objectives. Therefore, it would appear that a desire to restrict entry

as well as to reduce informational asymmetries motivated the re-emergence of medical

licensing laws in turn of the century America.

VI. Conclusion

This paper attributes the rise of university-educated professionals in late

nineteenth and early twentieth century America to the expansion of specialized

knowledge. University-trained professionals became important in many occupations at

the turn of the century because the growth of specialized knowledge made it

24

advantageous for individuals in certain fields to acquire a systematic advanced education.

Hence, the university gradually displaced the apprenticeship system as the primary mode

of vocational training in many professions. The early twentieth century became the age of

the “expert” largely because specialization and the growth of specialized knowledge

made it impossible to remain a successful generalist.

This paper also argues that the forces of specialization were an important factor

behind the emergence of occupational licensing laws. Urbanization, the rise of

impersonal exchange, and the growth of specialized knowledge exacerbated the problem

of asymmetric information by making it harder for simple reputation mechanisms to work

effectively, and by making it increasingly difficult for consumers to judge the quality of

professional services. By helping to eliminate “lemons” from the market, occupational

licensing laws may have functioned as a partial solution to this asymmetric information

problem.

Much scholarship on the emergence of occupational licensing focuses on its role

as an entry restriction. To test this hypothesis, we took advantage of cross-state variation

in the timing of licensing laws for a variety of occupations to determine the effect of

these laws on entry. For very few occupations was the evidence consistent with the entry

barrier hypothesis. Unfortunately, data limitations make it extremely difficult to

systematically test our alternative hypothesis. While we do furnish some systematic

evidence showing that for certain professions the likelihood that licensing was adopted

was positively related with the extent of the asymmetric information problem, our

econometric tests are not as decisive as we would like. Hence, we also examined the

history of the medical profession in a quest for qualitative evidence that might allow us to

more cleanly distinguish between the two main hypotheses.

Our study of the medical profession suggests that physicians in America have

historically used licensing in conjunction with private mechanisms in their efforts to

solve the problem of asymmetric information. During the colonial and early antebellum

periods, licensing, along with membership in local medical societies, played a role in

reducing informational asymmetries but did little to deter entry into the profession. As

the nineteenth century progressed, the emergence of proprietary schools and the

fragmentation of the medical profession eroded the value of medical licensing as a

25

solution to the asymmetric information problem. Hence, physicians lost interest in

licensing during the Jacksonian era, and in most states, medical licensing statutes were

repealed. When licensing re-emerged in the late 1800s and early 1900s, it was used to

eliminate low quality medical schools from the market, which eventually gave physicians

control over entry into the profession. The historical evidence therefore suggests that a

desire to control entry and solve the lemons problem were both important factors driving

the adoption of medical licensing laws in the late 1800s and early 1900s. Indeed, since

laws that solve asymmetric information problems will often also function as entry

restrictions that allow those producers who remain in the market to earn monopoly rents,

we believe that the history of medical licensing in America illustrates a fundamental

dilemma in the regulation of professional occupations.

References Akerlof, George (1970). “The Market for ‘Lemons’: Quality Uncertainty and the Market Mechanism.” Quarterly Journal of Economics 84: 488-500. Allison, Paul D. (1984). Event History Analysis: Regression for Longitudinal Event Data. Beverly Hills: Sage Publications. Calvert, Monte A. (1967). The Mechanical Engineer in America, 1830-1910: Professional Cultures in Conflict. Baltimore: Johns Hopkins University Press. Council of State Governments (1952). Occupational Licensing Legislation in the States. Chicago: Council of State Governments. Council on Medical Education (1920). “Medical Education in the United States.” Journal of the American Medical Association, 76: 379-388. Edwards, Alba M. (1942). Comparative Occupation Statistics for the United States, 1870 to 1940. Washington, DC: GPO. Flexner, Abraham (1910). Medical Education in the United States and Canada: A Report to the Carnegie Foundation for the Advancement of Teaching. New York: Carnegie Foundation. Fowler, Maureen J. (1966). Guides to Scientific Periodicals. London: The Library Association.

26

Friedman, Milton (1962). Capitalism and Freedom. Chicago: University of Chicago Press. _______________ and Simon Kuznets (1945). Income from Independent Professional Practice. New York: National Bureau of Economic Research. Gascoigne, Robert M (1985). A Historical Catalogue of Scientific Periodicals, 1665-1900. New York: Garland Publishing. Kessel, Reuben (1970). “The AMA and the Supply of Physicians.” Law and Contemporary Problems 35: 267-283. ______________ (1972). “Higher Education and the Nation’s Health: A Review of the Carnegie Commission Report on Medical Education.” Journal of Law and Economics 15: 115-127. Kett, Joseph (1968). The Formation of the American Medical Profession: The Role of Institutions, 1780-1860. New Haven: Yale University Press. Kim, Sukkoo (1995). “The Expansion of Markets and the Geographic Distribution of Economic Activities: Trends in US Regional Manufacturing Structure, 1860-1987.” Quarterly Journal of Economics 110: 881-908. ___________ (1998). “Economic Integration and Convergence: US Regions, 1840-1987.” Journal of Economic History 58: 659-683. ___________ (2000). “Urban Development in the United States, 1690-1990.” Southern Economic Journal 66: 855-880. Kuznets, Simon and Dorothy S. Thomas (eds.) (1965). Population Redistribution and Economic Growth: United States 1870-1950. Philadelphia: American Philosophical Society. Law, Marc T. (2003). “The Origins of State Pure Food Regulation.” Journal of Economic History, Forthcoming. Leland, Hayne (1979). “Quacks, Lemons and Licensing: A Theory of Minimum Quality Standards.” Journal of Political Economy 87: 1328-1346. Ludmerer, Kenneth M. (1985). Learning to Heal: The Development of American Medical Education. New York: Basic Books. Maurizi, Alex (1974). “Occupational Licensing and the Public Interest.” Journal of Political Economy 82: 399-413.

27

Merriam-Webster Inc. (1997). Merriam-Webster Dictionary. Springfield, MA: Merriam-Webster. Moore, Thomas (1964). “The Purpose of Licensing.” Journal of Law and Economics 4: 93-117. Price, Derek J. de Solla (1963). Little Science, Big Science. New York: Columbia University Press. Rothstein, William (1972). American Physicians of the Nineteenth Century: From Sects to Science. Baltimore: Johns Hopkins University Press. Shyrock, Richard (1960). “The College of Physicians in Philadelphia in Historical Perspective.” Transactions and Studies of the College of Physicians in Philadelphia. Society for the Promotion of Engineering Education (1930). Report of the Investigation of Engineering Education: Volume I. Lancaster: Lancaster Press. Starr, Paul (1982). The Social Transformation of American Medicine: The Rise of a Sovereign Profession and the Making of a Vast Industry. New York: Basic Books. Stevens, Robert (1983). Law School: Legal Education in America from the 1850s to the 1980s. Chapel Hill: University of North Carolina Press. Stigler, George (1971). “The Theory of Economic Regulation.” Bell Journal of Economics and Management Science. 2: 3-21. US Department of Commerce (various years). Census of Population. Washington, DC: GPO. Wallis, John J. and Douglass C. North (1986). “Measuring the Transaction Sector of the American Economy, 1870-1970.” In Engerman, Stanley and Robert Gallman (eds.). Long Term Factors in American Economic Growth. Chicago: University of Chicago Press, pp. 85-148.

28

Table 1: Specialization among scientific periodicals, 1780-1900 A. Number of periodicals by scientific field

1780 1800 1820 1840 1860 1880 1900 General science 27 34 45 69 124 181 218 Astronomy 0 1 1 3 6 9 23 Botany 0 2 3 11 22 33 59 Chemistry 1 7 7 14 18 25 48 Experimental biology 0 1 1 2 11 21 49 Geography 0 0 0 5 12 20 28 Geology 0 2 6 15 28 46 79 Mathematics 0 0 1 3 8 28 45 Natural history 2 3 5 20 40 60 72 Physics 0 1 1 3 4 11 25 Zoology 0 0 2 14 27 52 109 Other 0 0 0 1 4 19 38 Total 30 51 72 160 304 505 793 B. Percentage of periodicals by scientific field

1780 1800 1820 1840 1860 1880 1900 General science 90.0 66.6 62.5 43.1 40.8 35.8 27.5 Astronomy 0 2.0 1.4 1.9 2.0 1.8 2.9 Botany 0 3.9 4.2 6.8 7.2 6.5 7.4 Chemistry 3.3 13.7 9.7 8.8 6.0 5.0 6.0 Experimental biology 0 2.0 1.4 1.2 3.6 4.1 6.2 Geography 0 0 0 3.1 3.9 4.0 3.5 Geology 0 3.9 8.3 9.4 9.2 9.1 10.0 Mathematics 0 0 1.4 1.9 2.6 5.5 5.7 Natural history 6.7 5.9 6.9 12.5 13.2 11.9 9.1 Physics 0 2.0 1.4 1.9 1.3 2.2 3.2 Zoology 0 0 2.8 8.8 8.9 10.3 13.7 Other 0 0 0 0.6 1.3 3.8 4.8 Total 100 100 100 100 100 100 100 Source: Adopted from the tables contained in Part 3 of Gascoigne (1985). In this study, Gascoigne categorized a sample of roughly 900 scientific periodicals published in Germany, Great Britain, France, Italy, and the United States between 1670 and 1900 according to their field of specialization.

29

Table 2: Mid-twentieth century specialization in scientific knowledge A. Pure scientific fields

Approximate number of periodicals in field

Year Source

Astronomy 650 periodicals "not known to have ceased

publication"

1958 Crane, Louis F. (1958). An International Bibliography of Current Astronomical Serials. Ann Arbor: University of Michigan Press

Biology

5,000 “current titles” 1960 "Biological Abstracts List of Serials,” Biological Abstracts, 35: 4703-4742

Chemistry 1,560 "current and discontinued titles"

1960 Chemical Society (1960). Periodicals in the Chemical Society Library. London

Geography

657 “current titles” 1961 Royal Geographical Society (1961). Current Geographical Periodicals: A Hand-list and Subject Index of Current Periodicals in the Library of the RGS. London: RGS

Mathematics

800 titles circa 1965 American Mathematical Society, Mathematical Reviews.

Physics

850 “current titles” 1950 Bray, Robert S (1950). List of Periodicals of Physics Interest. Washington DC: Office of Technical Services

B. Scientific professions

Approximate number of periodicals in field

Year Source

Dentistry 1,255 "current and discontinued titles"

1962 Schmidt, Hans Joachim (1962). List of Dental Periodicals. Stuttgart-Degerlock: Verlag der Deutschen Doukumentenstelle für zahnärztliches Schrifttum.

Engineering

553 "current titles" 1953 Association of College and Reference Libraries (1953). A Recommended List of Basic Periodicals in Engineering and Engineering Sciences. Chicago: ACRL.

Medicine

2,700 "current titles" circa 1965 Index Medicus, Washington, DC: Public Health Service

Pharmacy

935 "current and discontinued periodicals"

1963 Andrews, Theodora (1963). World List of Pharmacy Periodicals. Washington, DC: American Society of Hospital Pharmacists.

Veterinary Medicine

570 "current titles" 1961 Commonwealth Bureau of Animal Health (1961). “List of Publications Searched by the Commonwealth Bureau of Animal Health.” Veterinary Bulletin 31: i-ix.

Source: Information in this table is taken from Fowler, Maureen J. (1966).

30

Table 3: Institutions of higher education: total number, enrollment and degrees granted, 1870-1970

1870 1890 1910 1930 1950 1970 Total number of institutions of higher education

563

998

951

1,409

1,863

2,525

Enrollment in institutions of higher education

(a) Total number (in thousands)

52

157

355

1,101

2,281

7,920 (b) As a percentage of the 18-24 years old population

1.1

1.8

2.9

7.2

14.2

32.1

Degrees granted from institutions of higher education

(a) Total number (in thousands)(b) Per 1,000 persons 23 years

old

9

n/a

17

n/a

40

20

140

57

497

182

1,065

223 Source: Series H 689, H 700, H 701, H 751 and H 755 from US Department of Commerce (1976)

31

Table 4: The timing of licensing regulations: number of states enacting licensing regulation by 10 year interval for selected occupations, 1870-1950

Profession Pre 1870 1871-80 1881-90 1891-1900 1901-10 1911-20 1921-30 1931-40 1941-50 Timing Unknown TotalAccountants 3 18 25 1 1 48Architects

1 6 11 12 10 6 46Attorneys 10 1 1 4 5 2 5 4 16 48Barbers 2 11 1 12 15 1 4 46Beauticians 1 20 21 3 45Chiropractors

1 17 21 4 1 44

Dentists 2 4 21 10 5 4 1 1 48Electricians 1 2 2 4 9Engineers 1 10 14 15 8 48Insurance Brokers

1 1 3 8 1 3 5 22

Midwives 4 2 5 1 1 1 14Practical Nurses 5 2 1 1 15 8 32Registered Nurses 23 20 4 1 48 Optometrists 15 21 5 1 6 48Osteopaths 8 19 8 3 1 8 47Pharmacists 7 24 10 4 1 2 48Physical therapists 1 1 2 5 9Physicians 5 11

18 7 2 5 48

Plumbers 2 2 3 6 3 16Psychologists 3 3Real Estate Brokers

6 23 6 5 40

Surveyors 2 4 3 3 1 20 33Teachers 10 1 2 2 2 4 1 26 48Veterinarians 2 7 15 11 3 10 48 Source: Adopted from Table A in Council of State Governments (1952).

Table 5: Fixed-effect regression estimates of the effect of state licensing laws on the number of workers per 1,000 persons in eleven occupations Architects

per 1,000 (1870-1930)

Architects per 1,000

(1880, 1900, 1920, 1930)

Attorneys per 1,000

(1870-1930)

Attorneys per 1,000

(1880, 1900, 1920, 1930)

Barbers per 1,000

(1870-1930)

Barbers per 1,000

(1880, 1900, 1920, 1930)

Constant

-0.042* (0.022)

-0.077*** (0.028)

1.74*** (0.37)

1.24*** (0.29)

1.01*** (0.38)

0.20 (0.29)

Percent Urban

0.36*** (0.079)

0.14* (0.070)

0.32 (0.70)

0.21*** (0.054)

1.71*** (0.41)

3.12*** (0.50)

Income per capita

0.000065** (0.00024)

0.00061** (0.00022)

0.00052*** (0.00018)

Regulation Dummy

-0.0051 (0.012)

-0.0030 (0.011)

0.14 (0.11)

0.064 (0.12)

-0.084 (0.056)

-0.067 (0.078)

Adj-R2 0.77 0.85 0.72 0.58 0.89 0.80 N 313 185 178 109 316 181 Beauticians

per 1,000 (1870-1930)

Beauticians per 1,000

(1880, 1900, 1920, 1930)

Dentists per 1,000

(1870-1930)

Dentists per 1,000

(1880, 1900, 1920, 1930)

Engineers per 1,000

(1870-1930)

Engineers per 1,000

(1880, 1900, 1920, 1930)

Constant

-0.036 (0.056)

-0.11 (0.14)

0.12*** (0.032)

0.14*** (0.049)

1.17*** (0.31)

0.38 (0.31)

Percent Urban

0.41*** (0.12)

0.22 (0.21)

0.20 (0.16)

0.11 (0.14)

1.41 (0.95)

0.92 (0.66)

Income per capita

0.00013 (0.000092)

0.000069 (0.000041)

0.00057** (0.00022)

Regulation dummy

0.15* (0.071)

0.18** (0.072)

-0.0016 (0.030)

-0.035 (0.021)

0.66 (0.16)

0.21 (0.13)

Adj-R2 0.84 0.82 0.70 0.80 0.72 0.77 N 326 189 322 189 322 189

Table 5 continued: Fixed-effect regression estimates of the impact of state licensing laws on the number of workers per 1,000 in eleven occupations Nurses per

1,000 (1870-1930)

Nurses per 1,000

(1880, 1900, 1920, 1930)

Physicians per 1,000

(1870-1930)

Physicians per 1,000

(1880, 1900, 1920, 1930)

Plumbers per 1,000

(1870-1930)

Plumbers per 1,000

(1880, 1900, 1920, 1930)

Constant

-0.45*** (0.12)

-0.54*** (0.19)

1.96** (0.33)

2.00** (0.60)

-0.047 (0.26)

-0.092 (0.43)

Percent Urban

1.60*** (0.29)

1.81*** (0.46)

0.94 (0.71)

0.70 (1.32)

2.40*** (0.42)

1.60*** (0.61)

Income per capita

-0.00011 (0.00016)

0.000080 (0.00032)

0.00044* (0.00025)

Regulation dummy

0.038 (0.10)

-0.092 (0.23)

-0.41 (0.31)

-0.54 (0.51)

0.11 (0.12)

0.091 (0.15)

Adj-R2 0.83 0.81 0.065 0.10 0.85 0.85 N 334 189 330 189 316 181 Teachers per

1,000 (1870-1930)

Teachers per, 1000

(1880, 1900, 1920, 1930)

Veterinarians per 1,000

(1880-1930)

Veterinarians per 1,000

(1880, 1900, 1920, 1930)

Constant

3.22*** (0.94)

4.01*** (1.23)

0.046** (0.019)

0.090*** (0.029)

Percent Urban

-2.93 (2.64)

-3.21 (2.62)

0.048 (0.046)

0.24 (0.80)

Income per capita

0.0034*** (0.00076)

0.000058*** (0.000046)

Regulation dummy

0.66* (0.30)

0.72* (0.37)

0.021 (0.082)

0.046 (0.041)

Adj-R2 0.88 0.90 0.74 0.66 N 154 88 269 158 Notes: Heteroskedasticity-consistent standard errors are reported in parentheses. Statistical significance at the 10 percent, 5 percent, and 1 percent levels are indicated by *, ** and *** respectively. Note that the number of observations differs across the professions because of the Council of State Governments (1952) did not provide complete data on the timing of occupational licensing laws for each state. Data on urbanization were taken from US Department of Commerce (1976). Data on income per capita were taken from Kuznets and Thomas (1965).

34

Table 6: Fixed-effect regression estimates of the effect of state licensing laws on the growth rate of workers in eleven different professions. %∆Architects

%∆Attorneys %∆Barbers %∆Beauticians %∆Dentists

Constant

-0.41 (0.28)

-0.20 (0.24)

-0.35 (0.30)

1.21*** (0.36)

0.35*** (0.13)

Population growth rate

1.77*** (0.028)

1.21*** (0.018)

1.11** (0.18)

0.94** (0.40)

0.98*** (0.097)

Regulation dummy

-0.11 (0.08)

0.10 (0.14)

-0.038 (0.12)

0.50** (0.20)

-0.14 (0.09)

Adj-R2 0.53 0.28 0.16 0.23 0.37 N 262 125 270 254 274 %∆Engineers

%∆Nurses %∆Physicians %∆Plumbers %∆Teachers %∆Vets

Constant

-0.23 (0.30)

0.50* (0.25)

-0.08 (0.16)

-0.15 (0.26)

0.60*** (0.14)

-0.037 (0.30)

Population growth rate

1.13*** (0.20)

0.66** (0.40)

0.98*** (0.11)

1.80*** (0.24)

1.073*** (0.14)

1.00*** (0.23)

Regulation dummy

-0.15 (0.10)

0.015 (0.10)

-0.17* (0.10)

-0.0092 (0.12)

0.018 (0.038)

-0.16* (0.09)

Adj-R2 0.61 0.74 0.38 0.65 0.87 0.54 N 274 210 243 265 132 224 Notes: Heteroskedasticity-consistent standard errors are reported in parentheses. Statistical significance at the 10 percent, 5 percent, and 1 percent levels are indicated by *, ** and *** respectively. Note that the number of observations differs across the professions because of the Council of State Governments (1952) did not provide complete data on the timing of occupational licensing laws for each state.

35

Table 7: Discrete time hazard model estimates of the factors influencing the adoption of licensing legislation for eleven occupations Architects

Architects Attorneys Attorneys Barbers Barbers

Constant

-3.01*** (0.39)

-3.00*** (0.40)

-1.55* (0.82)

-2.17* (1.20)

-8.31*** (1.81)

-9.01*** (2.24)

Percent urban

1.74* (1.00)

1.69* (1.00)

2.53** (1.05)

2.31** (1.01)

-0.84 (0.96)

0.17 (1.49)

Workers per thousand

2.20 (5.16)

2.61 (5.11)

-0.88 (1.10)

-0.96 (1.16)

6.17*** (0.96)

4.60* (2.71)

(Workers per thousand)2

1.34 (9.56)

0.94 (9.83)

0.24 (0.30)

0.29 (0.29)

-1.25** (0.58)

-1.11 (0.74)

Time varying hazard rate

No Yes No Yes No Yes

Log-likelihood -93.13 -93.04 -38.56 -35.41 -76.39 -56.73 N 284 284 83 83 286 286 Beauticians Beauticians Dentists Dentists Engineers Engineers

Constant

-6.20*** (1.07)

-6.26*** (1.14)

2.93*** (0.63)

-3.44*** (0.66)

-3.67*** (0.47)

-5.32*** (1.25)

Percent urban

-2.57 (1.72)

-1.65 (1.97)

0.98 (1.01)

1.34 (1.08)

-0.48 (0.89)

-2.49 (1.77)

Workers per thousand

13.32*** (2.61)

8.25** (3.95)

7.10*** (2.52)

5.33* (3.18)

1.51*** (0.93)

0.33 (0.93)

(Workers per thousand)2

-5.43*** (1.83)

-3.37 (2.13)

-3.62** (1.54)

-3.15* (1.88)

-0.16* (0.09)

-0.07 (0.15)

Time varying hazard

No Yes No Yes No Yes

Log-likelihood -37.48 -35.63 -86.93 -73.53 -75.13 -50.59 N 329 329 162 162 312 312

36

Table 7 continued: Discrete time hazard model estimates of the factors influencing the adoption of licensing legislation for eleven different occupations Nurses Nurses Physicians Physicians Plumbers Plumbers

Constant

-4.39*** (0.48)

-4.91*** (0.20)

-6.11** (2.54)

3.49*** (0.78)

-6.23*** (1.16)

-6.92*** (1.07)

Percent urban

1.28 (0.95)

1.47 (1.18)

1.53* (0.83)

1.49* (0.85)

-1.63 (1.46)

0.22 (1.66)

Workers per thousand

4.43*** (0.56)

3.73*** (0.73)

6.12** (3.09)

1.41** (0.55)

3.38** (1.36)

2.31 (1.86)

(Workers per thousand)2

-1.25*** (0.24)

-0.99** (0.31)

-1.87** (0.91)

-0.07** (0.03)

-0.67 (0.44)

-0.53 (0.50)

Time varying hazard

No Yes No Yes No Yes

Log-likelihood -89.09 -64.73 -89.44 -77.07 -23.11 -21.22 N 268 268 162 162 308 308 Teachers

Teachers Veterinarians Veterinarians

Constant

-6.85** (3.07)

-5.95** (2.95)

-3.67*** (0.56)

-3.08*** (0.64)

Percent urban

0.48 (2.10)

-0.08 (2.25)

3.32*** (1.10)

3.69*** (1.30)

Workers per thousand

1.34 (0.95)

1.06 (1.10)

11.94 (8.32)

20.68** (10.50)

(Workers per thousand)2

-0.06 (0.09)

-0.05 (0.09)

-1.51 (2.76)

-36.39 (30.39)

Time varying hazard

No Yes No Yes

Log-likelihood -23.35 -21.33 -83.03 -67.89 N 57 57 197 197 Notes: The coefficient estimates show the change in the log-odds of adopting licensing legislation for a given occupation resulting from a one-unit change in each independent variable. Statistical significance at the 10, 5 and 1 percent levels is denoted by *, **, and *** respectively. Heteroskedasticity-robust standard errors are reported in parenthesis.

37