Lentis/Path Dependence

The original QWERTY keyboard layout, designed Latham Sholes in 1878, is optimized for typewriters but not for modern computers.

Path Dependence is a socioeconomic theory. It has long been theorized that human beings, though individually imperfect, seek en masse to find the most efficient solution to any problem. Through exploration and competition they find the best, most beneficial arrangements for themselves, and thus, with time, society optimizes itself. This is a very neat, tidy theory - the sort of theory that pleases the mathematically minded. However, it fails to account for reality. Anyone who makes an examination of the world they find themselves in is sure to find inefficiency rampant; inefficiency not only for them, but for everyone, and often with a clear, or even obvious solution. Path dependence provides an explanation for this discrepancy.

The Theory of Path Dependence

In traditional economic theory, only a single, optimal equilibrium point exists in any market.

Path dependence suggests the existence of multiple equilibria, each the result of a different set of paths.

Path dependence was discovered and named by economists, as an explanation for the pattern by which new products or ideas entering a marketplace reach acceptance. It was developed in response to neoclassical economics, which theorized that in any market there was a single point of equilibrium defined by the optimal solution to market variables (i.e. the solution which yielded the greatest benefit to the majority), and that this point of equilibrium would always be reached regardless of the path by which it was found. This theory fails to account for real-world situations wherein the market equilibrium reached does not represent an optimal solution. The theory of path dependency posits that most markets have multiple equilibria, none necessarily optimal and some preferable to others, and that the equilibrium reached through market adoption is dependent on the particular path that market adoption takes. In other words, noise and minor variations have a significant effect on the outcome of any market.

The mechanisms which drive path dependance are switching costs and forking costs. Where these two mechanisms do not exist, path dependence will have no effect. Once path dependence comes into effect, it often remains until some new development shifts the points of equilibrium completely, until an alternate path becomes sufficiently attractive to outweigh switching or forking costs, or occasionally until it is defeated by intelligent effort.

Switching Costs

A switching cost is any cost associated with moving from one path to another. This can take the form of a literal monetary cost, such as new investment or loss of old investment, or a time cost, or it can be a less tangible cost. In essence, any inconvenience associated with a change of market paths represents a switching cost, including such things as confusion or discomfort. The key property of a significant switching cost is its immediacy. In many cases, a sub-optimal path or equilibrium will itself carry a significant cost (e.g. the time and comfort lost to typing inefficiently), but this cost will be accrued over time. A switching cost is significantly higher than the cost incurred by the switched-from path at any given point in time. If path-determining decisions are irreversible, there may effectively be an infinite switching cost.

Forking Costs

A forking cost is a cost which prevents different market actors from adopting different paths. In some situations, an optimal path may be more costly than a less-optimal one simply because a majority of the market has adopted the less-optimal option. In cases where the path is chosen by a regulatory body, this cost may come in the form of whatever enforcement power that body has. Otherwise, it often takes the form of inconvenience or higher costs for all paths. Format wars provide an excellent example: multiple paths cause significant inconvenience to manufacturers, who must build multiple kinds of devices and media in order to access the entire market, as well as users, who must buy multiple kinds of devices, and who are often hindered by incompatibilities. The effect of a forking cost is that a single path is eventually chosen. Usually, this is the most popular path rather than the optimal one. Often, when a new technology enters a market, a number of paths are explored, as forking and switching costs have not yet manifested. However, once a single path gains a critical mass the cost of remaining on another path becomes prohibitive.

Beyond Economics

Although path dependence was originally developed as an economic theory, it is conjectured that it may be a basic rule by which many natural systems operate. Regardless, path dependence may clearly be seen outside the explicit markets of economics. Law is thought to be a path dependent system , and it has been suggested that evolution and history may be path-dependent.

Examples of Path Dependence

QWERTY and Dvorak Keyboards

The standard Dvorak keyboard, widely considered more efficient than QWERTY.

The QWERTY keyboard was designed and optimized for Latin script typewriters, and sold to Remington in 1873. The primary design concern was the prevention of typewriter jams, achieved by slowing typing speeds and separating frequently used key combinations. The Dvorak keyboard was patented in 1936. It is claimed by proponents to require less finger motion, yielding an increase in typing rate and a reduction in errors compared to the QWERTY keyboard. However, QWERTY is still the standard keyboard layout even though a more optimized layout has been available since 1936. While Dvorak users may type more efficiently individually and on their own equipment, they face a serious impairment when required to use the majority of keyboards on the market.

Metrication in the United States

This measuring cup displays the metric units first to the majority of users, who are right handed. This is an example of an attempt at metrication in the United States, which failed because users can simply turn the cup around.

The metric system dates to the 16th century and is the official system of measurement for most of the world. The United States is the only industrialized country that has not adopted the metric system, although it has attempted to switch in the past. Most recently, the 94th Congress passed The Metric Conversion Act in 1975. The conversion was voluntary and promoted through public education, with a planned transition period of 10 years. However, the public was not receptive , and ultimately metrication in the United States failed. Because the United States made metrication voluntary, the original path remained dominant due to the lack of switching costs. The long transition period may also have been a mistake. When a new path is chosen, there must be a dedicated transition point. The long transition time prevented a coordinated, synchronized transition, resulting in apathy and a failure to change paths.

x86 Processors

The original Intel 8086, which set the path for x86.

Intel released the original Intel 8086 16-bit CPU in 1978. In the subsequent decades, the majority of Intel processors built upon the instruction set that the 8086 introduced, while remaining backward compatible with it. The x86 instruction set advanced from 16-bit to 32-bit in 1985 with the release of IA-32, and eventually came to dominate the consumer PC market. When the time came to transition to 64-bit, Intel believed that the firm path set by backwards compatibility had become a liability, and that a new architecture would be more efficient. On October 4, 1999 , the Itanium was announced. The Itanium failed quickly in the marketplace due to poor or nonexistent backwards compatibility, among other concerns, and became the subject of much derision, gaining monikers such as "Itanic" . AMD, perhaps realizing the power of path dependence, announced its own 64-bit version of the x86 architecture called AMD64, shortly after Itanium in 1999 . This architecture was successful because it allowed users to continue on the original path without switching costs. In fact, it was so successful that it forced Intel to release its own 64-bit version of the x86 architecture in order to remain competitive. Thus, Intel was forced back into its previous, non-optimal path.

Examples of Path Dependence Overcome

Railroad Track Width in the Southern United States

Graphic list of track gauges

During the 19th century, there were a number of different track gauges in the United States. Although a common gauge would have greatly improved the efficiency of rail transport, any effort to establish one would hinder the competitive interests of the various railroad companies. However, as the South began its recovery from the war, it became evident that economic reconstruction would require trade with the rest of the country - an impossibility so long as differences in rail gauge existed.

In 1884 the Illinois Central Railroad, which operated in both the North and the South, found it necessary to begin changing the gauge of its Southern lines to conform to the Northern width. Though it did so at great expense, it gained an advantage, forcing other railroads to match its effort in order to remain competitive.

On February 2-3, 1886, operating officers of the South's railroads met at the Kimball House in Atlanta in a "Convention ...called for the purpose of fixing date and arranging details for change of gauge." A gauge of 4 feet, 8 and a half inches was selected, and on May 31st, 1886, thousands of workers across the southern United States manually adjusted 11500 miles of track to the new standard gauge, completing the work in 36 hours.

The reason this standardization happened is that the actual cost of keeping the gauges the way they were significantly outweighed the switching cost of standardization. In effect, the equilibrium point was shifted when trade with the North became necessary, but only a single company was capable of conducting it efficiently.

Swedish Road Orientation Change

Kungsgatan, Stockholm, on Dagen H

Prior to September 3rd, 1967, Swedish traffic laws dictated that cars drove on the left side of the road. The change from left to right was instigated by the fact that Sweden's neighbors drove on the right; if Sweden continued to drive on the left, it would result in a large number of lane switch-overs, and would thus present a significant safety hazard. Through meticulous preparation over the course of several years, Sweden prepared for the day that the lanes would switch, and when the day arrived, the plan was executed. In this instance, the forking cost and switching cost were in opposition, and the need to diverge from the current path overcame the switching cost.

Generalized Lessons

History Matters

The current environment is the product of past decisions. The decisions we make now will influence the environment of the future. History matters because paths taken at critical junctures may be irreversible. Therefore, all decisions must be made with a mindfulness of future implications and must not be based solely on contemporary conditions.

Overcoming Path Dependence

Once reached, sub-optimal equilibria will resist all attempts at change, due to the strength of path dependence. If a point arrives at which it has become more optimal to detach from the paths of the past, success requires a single, massive, non-voluntary effort. Choices must be made with an understanding of the difficulty in reversing them, should they turn out to be poor.

Future Work

There are many more instances of path dependence and lessons to be taken from them. Future work is necessary in order to understand how path dependent sub-optimal equilibria can be best recognized. Increased recognition and understanding will lead to increases in efficiency across the human experience.