Shakeouts in the Music Industry

The history of music industry concentration and the chronology of events provide general evidence against technology always being the direct cause of shakeouts. At the beginning of the industry’s life (1890-1900), there were three major firms producing the vast majority of audio products: Victor, Columbia, and Edison. This included both the machines- cylinder and record players- and the actual cylinders and records. Patents on these machines were a major barrier to entry, but major innovations from 1900-1910 and the expiration of important patents in 1914 resulted in industry deconcentration. Early record production required live-action recording to produce each record, requiring either multiple record writers present during a performance or multiple performances. From 1914 to 1919, the number of firms manufacturing records and record players grew on average by 44 percent annually. Demand was stimulated as a result of a new variety and quantity of available products on the market, and the period was characterized by heavy innovation in the music, particularly by small producers. However, from 1919 to 1925, the number of producers declined at an average annual rate of 14.4 percent. Larger firms were able to capitalize on the small producers’ innovations, resulting in imitation as well as several horizontal mergers. The onset of the Great Depression and World War II finalized the reconcentration of the music industry. Prior to 1948, Columbia, Decca, RCA Victor, and Capitol were responsible for three-fourths of record sales in America.

Following the war, a new innovation reshaped the industry: magnetic tape recordings. Previously, records were produced in a very tedious and unforgiving fashion. Errors in the performance for a recording would require the artists to execute the piece perfectly – start to finish – in order for the recording to be successful, but magnetic tape

allowed a particular section with an error to be spliced out and replaced by a re-recorded part. Magnetic tape machines were also much cheaper. By reducing the amount of studio time required and also lowering the costs of starting up a recording business, magnetic tape technology was followed by an increase in the number of companies producing LP (long-play) records from eleven to two thousand between 1949 and 1954 (Gelatt 1954).

By 1956 independent firms held around 52 percent of the music recording industry’s total market share, increasing to the industry’s peak in 1962, at which time independent firms accounted for 75 percent. Afterward, major firms began to reacquire market share, primarily through horizontal mergers, and the number of firms in the industry began to shrink.

Why did music industry shakeouts happen?

This prompts us to seek an alternative explanation to technological changes for the causes of the most recent extended music industry shakeout (1962-). Several technological improvements turned out to be exogenous (allowing universal adaptation) rather than endogenous (proprietary and thus concentration-inducing). The nature of the technologies Alexander cites tended to be scale-reducing, thus reducing barriers to entry. Developments in musical technology over the past 50 years have been consistently scale-reducing, though the trend for a large portion of that period has been toward consolidation. Magnetic tape and compact disc players became commercial and low-cost home appliances, and their respective means of creation grew as common (tape recorders, CD-burners, etc.). Computer-based music recording and playback has become more widespread. Still, the number of firms has been decreasing. Currently, the music market is dominated by six major firms: Time/Warner, Sony/CBS, Thorn/EMI, Philips-Polygram/PMG, Bertelsmann Music Group/BMG, and Matsushita/MCA.

One important factor stands above all other explanations for this consolidation: distribution. While prior to 1962 there were several strong and independent music distributors who provided an alternative to the major firms’ distribution networks, major firms began making significant buyouts in the 60s onward, creating a dominant market tendency toward the horizontal integration of distribution. Many independent distributors went bankrupt, and this tendency grew even more exaggerated in the 1980s. The six major firms mentioned above presently constitute almost the entirety of the industry’s market share at the distributor level.

In light of this evidence, one revised hypothesis is that technology can play a role in market concentration in as much as it augments scale economies. Technological innovations such as widespread personal computers with sound processing and recording capabilities, as well as advanced software for manipulating recordings, have reduced the necessary scale to begin producing consumable music recordings to anyone with or without talent, with just a $300 personal computer, a $30 microphone, and some small degree of sound engineering skills. The internet has also drastically reduced the scale required for significant levels of distribution, with peer-to-peer sharing networks, internet-based record stores, and social networking pages like MySpace.com.

On the other side of the story, some non-technological things may account for firm “lock-in” or other phenomena that lead to high industry concentration. Distribution strategy is one possible example of this, but it is likely that the dominance of particular firms that allowed them to construct their distribution networks shares a cause with their distribution strategies. Music is a very unique kind of product. Each new “product” (a song or album) also happens to be distinctly associated with a set of individuals. The quality of the music itself is controlled from a non-technological (in the physical sense) set of innovations relating to meter, pitch, tone, content, or overall theme. Some major firms may have the musical brainpower to “get it”- a group of experts, who manage bands and affect the musical product, that ultimately represents a stock of knowledge the firm has about stimulating and satisfying demand for music. Furthermore, labels fortunate enough to enlist legendary bands, perhaps by only good fortune, gain a long-lasting advantage, both from their experiences with a popular band (more concerts, albums, events, merchandising, etc.) as well as from the profits, which attract more expertise, which attracts and creates better bands, etc. There appear to be many opportunities for self-reinforcement in the industry. Overall, the technology-based shakeout story lacks explanatory power in music.

Source:

Alexander, Peter. New Technology and Market Structure:

Evidence from the Music Recording Industry. Journal of Cultural Economics, Volume 18, 113-123, 1994.

The post Industry Concentration and Shakeouts in the Music Industry appeared first on Rare Essays.

Using an optimal contracting model, Franco and Mitchell compare their model’s predictions when CNCs are allowed and when they are prohibited. They conclude some important things: the model explains the higher turnover in places where CNCs are prohibited; enforcement of CNCs in a region encourages greater firm numbers in early industry stages, especially where innovation is a key factor, thus explaining the early advantage of Route 128 which was eventually overcome by Silicon Valley; and we should expect to see in the data that concentrated industries seek CNC-enforcing areas, while competitive industries should be less likely to seek out such protection.

Their model, however, makes two key assumptions: first, that wages can’t be “backloaded”- in other words, employees can’t agree to be paid less than they would in their alternative option (to form a spin-out) for one period, and then get paid more to compensate in a later period; and second, that the level of the employee’s knowledge of the production process is known only by the employee. These may be generous assumptions that, when changed, could possibly alter Franco and Mitchell’s results drastically. One way to test their “backloading” assumption is by exploring the ways in which companies (especially in CNC-prohibiting regions) create economic incentives to depress labor mobility, and how often they do so. If wage backloading plays a significant role in those companies’ hiring practices, Franco and Mitchell’s model may be leaving out an essential explanatory variable. Their information asymmetry assumption also requires testing. On one hand, its importance can be explored by relaxing it in their model and testing its implications; on the other, instances in which employers actually have information about what their employees know about the production process should also be helpful.

Substitutes for Non-Compete Contracts

If non-compete contracts are illegal, one alternative means of restricting labor mobility is by “backloading” wages. In order for a firm to keep its employees from moving to a competing firm, the firm may decide to backload the wages, often in the form of pensions, options, health insurance, and other benefits that could only be captured if the employee stays with the firm over a certain time period. Franco and Mitchell (2005) assume backloading as impossible in their model. Rebitzer (2006) overlooks it.

A non-compete contract can initially be helpful in the early stages of an industry. As described by Rebitzer (2006), if employees were to “hop” around to other firms, the likelihood that knowledge acquired in one firm would be employed in another firm increases. These knowledge spillovers can hurt innovation by reducing the rewards to investing in human capital. On the other hand, abolishing non-compete contracts can be more helpful to local firms in the long run. If non-compete contracts are unenforceable, their elimination can lead to more turnover and more competitive entrepreneurs, assisting local firms in competing with out of state industries. Comparing Silicon Valley to Massachusetts’ Route 128, Franco and Mitchell (2005) found that Massachusetts’ Route 128 was more productive initially, but was eventually overtaken by Silicon Valley.

If non-compete contracts were unenforceable in Silicon Valley, it would be interesting to see if firms in Silicon Valley tended to backload wages more than Massachusetts’ Route 128, since the non-compete contracts in Massachusetts’ Route 128 were enforceable. If this is the case, the results by Franco and Mitchell may not have come about because of different non-compete regulations.

Burdett and Coles (2003) model a similar scenario of non-compete contracts, but they instead use wage-tenure contracts that give employees an incentive to remain in the firm and not move to a competing firm. In their story, each firm offers a wage-tenure contract that implies any employee’s wage smoothly increases with tenure. We can also compare the tenure policies of firm in Silicon Valley and Massachusetts’ Route 128, and see if the results of the model by Burdett and Coles (2003) are consistent with the data.

Wage Structuring

Wage-based policies can also serve as an alternative to non-compete contracts for reducing labor mobility-based knowledge spillovers. During a training period, for example, a worker gets paid less than his outside option (moving to other firms). However, depending on the importance of the information and the enforceability of the non-compete contracts, promising satisfactorily high wages to workers after their training period will stop them from changing employers. This kind of policy seems to successfully prevent spillovers across companies in the industry, but it does not affect industry clustering or profitability. It may even be the case that these policies can be more efficient than enforcing a legal framework for non-compete contracts. Fosturi and Ronde (2002), in their study “High-tech clusters, technology spillovers, and trade secret laws” theoretically demonstrate that even though information spillovers caused by labor mobility are prevented, industry clusters and profitability remain undisturbed.

One avenue of investigation to pursue would be to find a data set including wages, labor mobility, and the density of cluster in a region to test the assumptions given above. There is an industry cluster of biotech companies in San Diego, which is also supported by educational system in the region; several education institutions provide education from the undergraduate to doctorate level in biotechnology, providing a local labor pool. There are a large number of high quality workers who are educated and trained by the companies, though labor mobility among firms is high in this industry.

Fringe Benefits

Different legal frameworks for labor mobility-reducing contracts in different states prompts a search for other strategies that might be undertaken by firms to prevent labor-related information leakages to competing firms. Facing a lack of legal framework (like non-compete contracts) that can prevent workers from quitting and working in competitor companies overnight, firms need to utilize economic incentives to reduce information spillovers. Though the spillovers have a second level benefit through clustering, they also may have first level costs manifested by forgone opportunities for new innovations, for example. For the workers, one of the costs of labor mobility is an income loss due to foregone fringe benefits (Mitchell, 1983).

Controlling for other variables like unionization (Freeman, 1981), market concentration, regulations setting minimum prices and restricting entry, and profit regulations (Long and Link, 1983), we find that a substantial amount of variation in individual wage and fringe benefits is accounted for by industry differences. Dickens and Katz (1987) argue that high wage industries have lower quit rates, high labor productivity, more educated workers, longer work weeks, a higher ratio of non-wage (fringe benefits) to wage compensation, high unionization rates, bigger initial sizes and bigger average size of firms, higher concentration ratios, and more profits. However, what the studies above do not address is the effect of specific industry characteristics on the endogenous choice of fringe benefit costs by firms. We would expect the industries with fast innovative processes – without a legal framework that restricts labor mobility -would reflect higher levels of fringe benefit-related costs, and likely have better-defined promotion structures that discourage labor turnover.

The post Labor Mobility and Industry Agglomeration in Silicon Valley appeared first on Rare Essays.

“Variety” in one form or another can be predictive of an individual’s choice to pursue self-employment, whether it is preference for variety, actual experience of variety, or a combination of both. Variety in experiences can manifest itself in different ways. Sources of knowledge about entrepreneurship can appear in family history, among friends, in education, in the media, etc. It can also appear, most significantly, in an individual’s employment history. However, variety of this kind as a predictor for choosing self-employment can lead to very confusing results if not integrated into the context in which it actually plays a causal role. Sometimes, taste for variety causes varied work experiences; other times, varied work experiences can cause taste for variety. In other cases, taste does not even enter the picture except in the obvious case of “taste for subsistence”: some people hold different jobs only by necessity.

In light of this, Van Praag and Van Ophem (1995) wisely draw a distinction between influences on self-employment based on willingness versus opportunity. Their model’s estimation suggests that many young Americans possess the willingness to switch for self-employment, but lack the opportunities (primarily capital) to switch. More generally, they find that entrepreneurial abilities that compensate for lack of capital are rare. While taste for variety can be represented by variety in prior work experience, this potentially confuses issues of willingness with issues of opportunity even before we confront the same problem with regard to entrepreneurial choice.

One way of sidestepping this confusion is by actually modeling individuals’ decision-making processes instead of externally tracking their data over time to discover predictors of their choices. Its parameters can be found by observing directly what an individual values, both by questioning and by measuring business and non-business variables. Opportunity-related variables (e.g., wealth, credit access) can be integrated later in order to answer the actual broader question of who becomes a functioning entrepreneur. First, however, variety in taste or experience among other things must be used to ascertain willingness (or even predict attempts) for self-employment.

Summary of Katz (1992) Psychosocial Cognitive Model

Focusing solely on the choice to become self-employed or not (a simpler explanatory objective), variety of at least some kind certainly makes a difference. Katz (1992) proposes a psychosocial cognitive model (PCM) of employment status choice. It utilizes individuals’ psychology (through values and decision-making processes) as well as personal history and social context as factors having an effect upon decision-making.

The individual’s decision process begins with some kind internal discovery or external change (a changed awareness or dissonance) interacting with that person’s values. “Push” or “pull” effects can effectively describe this process (as explained in Vesper (1990)). The values likely to trigger an employment decision making process include desires for autonomy, creativity, material gain and power, and social integration.

The decision process consists of considering employment alternatives. The main source of these opportunities is one’s memories, and the heuristic of availability is a statistical means of representing the likelihood of retrieving information about opportunities as a function of its presence in the individual’s memory. The information, in turn, is determined by past exposure to such information. Katz (1992) specifically uses family, education, peers, prior work, and cohorts (age, racial, gender, ethnic, and geographic) as sources of experience about employment alternatives.

Once a set of alternatives is developed, it is evaluated against the initial dissonance. If the set is satisfactory, the individual begins the process of selection. If it is unsatisfactory, the individual either searches for more possibilities from memory (repeating the availability cycle), or constructs new possibilities. The most likely form of construction is the generalization of past work experiences to new ones, such as a former computer repairman considering building and selling new computers out of his home. Another, less likely form of construction is the creation of novel alternatives that directly solve the dissonance that caused the initial search. For example, if a strong force that caused someone to be unemployed was inflexible work hours, an alternative containing a satisfactory work schedule will be included in the set of alternatives.

Finally, the development of the set of alternatives completes and the agent must choose a course of action. The representativeness heuristic, from Tversky and Kahneman (1974), underlies this process. It refers to the individual’s assessment of the likelihood that an alternative will lead to a preferred outcome (like financial success). While maximization of likelihood is generally preferred, qualitatively rational processes play a large role. Alternatives are chosen on the basis of factors such as how well they fit the agent’s values or how familiar they are. Following that initial decision, the implementation of the alternative is constrained by environmental factors which can affect whether action will be taken or not. This can be compensated for by a second round of PCM; while the first round explores what the individual wants to do, the second round explores how he wants to do it.

Katz (1981)

In order to attempt the model, Katz (1981) drew 17 variables from the Panel Study of Income Dynamics, some of which he utilized as proxies for general breadth of work experience- exposure to variety- and some which were proxies for more direct exposures to work experiences through the individual’s own work, through family, or through membership in groups with above-average tendencies toward self-employment. Incorporating variables such as these should lead to much more reliable results than using the traditional handful of experience surrogates.

The 17 variables are as follows:

1. Father’s Self-Employment

2. Father’s Education

3. Employment Status of Respondent’s First Job

4. Number of Different Jobs Held

5. Age is Young (16-30) or Old (55-98)

6. Gender is Male

7. Ethnicity in High Self-employment Incidence Group

8. Own Education Less Than High School

9. Exposure to Variety: Reads newspaper

10. Exposure to Variety: Watches Television

11. Exposure to Variety: Goes to Religious Services

12. Exposure to Variety: Goes to Social Clubs or Organizations

13. Exposure to Variety: Goes to Bars or Taverns

14. Exposure to Variety: Belongs to Labor Union

15. Exposure to Variety: Known by Name to Neighbors

16. Exposure to Variety: Relatives Within Walking Distance

17. Exposure to Variety: Farm/Small Town Childhood

Outcomes were categorized in 5 possibilities: changes from wage-or-salaried work to self-employment, from one wage-or-salaried job to another, from self-employment to wage-or-salaried work, and no-change in status for the wage-or-salaried or the self-employed. This variable set mostly focuses on a “first-round” mode of thinking about self-employment (“what would he do if he could?”). Nonetheless, broadly speaking, a PCM approach fundamentally copes better with the problems that arise from one’s willingness to become self-employed in conflict with his opportunity to do so. Instead of considering variables of willingness and opportunity separately and then combining them to form a model, the PCM could potentially assess the actual point at which these variables interact: within the cognitive processes of the individual.

This is reflected by the advantages to the model cited by Katz. A PCM approach strongly favors the incorporation of a vaster scope of qualitative data to accommodate consistent qualitative findings, such as the self-employment choices of children of self-employed parents. The model, by virtue of the fact that it avoids pure econometric analysis, also demonstrates robustness when provided with non-ideal data sets. Katz claims that after operationalizing about one-half of the Katz (1981) model, almost 42% of cases’ placements were predicted correctly (out of 5 choices, with which a random selection would yield 20% correct prediction). Ultimately, constructing a cognitive model that models decision-making can yield different and new insights compared to traditional, occupational tracking models.

Citations:

Jerome A. Katz. A psychosocial cognitive model of employment status choice. Entrepreneurship: Theory and Practice 17.n1 (Fall 1992): pp29(8).

C. Mirjam Van Praag, Hans Van Ophem (1995)
Determinants of Willingness and Opportunity to Start as an Entrepreneur
Kyklos 48 (4), 513–540.

The post Do Psychosocial-Cognitive Factors Explain Variety in Tastes and Experience? appeared first on Rare Essays.