Official Japanese Data

Compare Tokyo, obviously an urban area, with Aomori, a rural area. Table 1 presents data on items lost and recovery rates for the two areas in 2001.

Table 1. Lost Property in Tokyo and Aomori, 2001

Sources: Kōban from Sōmucho Reference TōkeikyokuTōkeikyoku (2000:188). Tokyo data from Tokyo Metropolitan Police, http://www.keishicho.metro.tokyo.jp/toukei/kaikei/kaikei.htm (2002). Aomori data from Aomori Prefectural Police, http://www.police.pref.aomori.jp/keimubu/otoshimono.htm (2001).

The urban/rural comparison reveals both interesting similarities and interesting differences. The percentages of objects and cash that are returned to owners are quite high, roughly the same, and near the national average.Footnote ⁴ In both cases, more than 60% of objects and more than 90% of cash are recovered by either the finder or the original owner. But note the disposition of found objects. Tokyo awards a much higher percentage to finders than Aomori, and Aomori returns a much higher percentage to original owners. Perhaps Tokyo finders are more aggressive in pursuing their claims, or perhaps Tokyo losers give up more easily.

This difference aside, the rest of the figures line up about as expected. The number one found object in both areas is umbrellas, with wallets coming in second and third. Bicycles turn up more often in Aomori, perhaps because of greater distances between public transportation facilities. Most interestingly, the disposition of cash is nearly identical. In roughly three-quarters of the cases of found cash, the owner turns up—and only a very small portion of the cash goes to neither the finder nor the owner.Footnote ⁵

To get a sense of whether Tokyo and Aomori data accurately represent national lost-and-found patterns over time, I examined national lost-and-found data for the period 1973–2000. Figure 1 details the number of Loss Reports and Finders' Reports for the period. Finders' Reports are filed much more frequently than Loss Reports, resulting in data that appear to show more finds than losses. Most basically, as the figure shows, loss rates and find rates have generally increased over the period (data unavailable 1982–1983), with a “bubble” coinciding with Japan's economic bubble in the late 1980s and early 1990s.

Figure 1. Lost and Found Reports, 1973–2000

To get a more complete picture, it is useful to divide the data into objects and cash. Figures 2 and 3 show the amount of cash and the numbers of items listed in the Loss Reports and Finders' Reports.

Figure 2. Lost and Found Cash, 1973–2000

Figure 3. Lost and Found Items, 1973–2000

These two charts suggest some interesting trends. First, note that reported cash finds as a percentage of reported cash losses have decreased significantly over the 25-year period (from nearly half in the mid-1970s to less than one-third in 1998). Again, multiple phenomena may account for the change; among other things, finders might be less honest than they were in the past, or losers might be reporting more cash losses.

Second, note that the opposite phenomenon has occurred for object finds. Reported object losses as a percentage of reported object finds have increased significantly (from less than half in the 1970s to more than three-fourths in the late 1990s). Perhaps fewer object losses are being reported (perhaps items may be more replaceable), or more object finds are being reported.

Third, note that the loss rate for cash exceeds the find rate, while the find rate for items exceeds the loss rate. In this connection, note also that the find rate and loss rate exhibit roughly the same trends in both charts. Although the report rates have changed over time, the same basic ups and downs appear. To check the relations more precisely, I compared the correlation of the first differences—the annual rates of change—between (1) cash lost reports and cash found reports and (2) object lost reports and object found reports. The correlation for cash was a highly significant 0.858 (p=0.000). The correlation for objects was only 0.283 (p=0.160), not even marginally significant. While the data may be imperfect, these results may indicate that the system is more effective in reuniting losers with cash than noncash items, as find report rates more closely track loss rates.

To further explore incentives, I examined the data on the disposition of found property, as detailed in Figure 4. As the figure indicates, owners have become more aggressive in retrieving both property and cash in the last 25 years, but cash has a higher return rate. Again, multiple explanations may be possible, including the easier identification of cash owners (cash is often lost in wallets that contain identification). In any event, the data suggest that while cash finds exceed reported cash losses, cash losses that are reported are more likely to be returned to the owner than object losses. The increase in the percentage of returned items over time may be indicative of an increase in the value of portable objects such as electronics in the past two decades.

Figure 4. Found Items Returned to Owner, 1973–2000

Comparison to the United States

Unlike Japan, the United States keeps no national statistics on lost-and-found property, and I was unable even to locate any state that maintains such data. Accordingly, I turned to city-specific data for insights, but note nevertheless that any unit of comparison is incomplete. In any two systems, and certainly in comparisons between Japan and the United States, there are likely to be differences in objects lost, the nature of places frequented, and conceptions of what objects people should turn in to authorities. Even the nature of objects carried differs—people in Japan tend to use cash, while people in the United States tend to use cards (Reference MannMann 2002)—and the data do not compensate for such factors.

Those caveats stated, the best (but imperfect) comparison in the United States with Tokyo, at least, is probably New York. New York's lost-property statute requires police to accept property, give receipts, and give notice of the finding if they “have reason to believe that a person has an interest in found property” (New York Personal Property Law §253 2003). As Part II will show, the New York statute is also somewhat similar to the Japanese statute in awarding the property to the finder after a specified period of time. But when I called the precincts specifically designated by the New York City Taxi and Limousine Commission to receive lost property, I was informed that those precincts only keep logs, not statistics. I filed a Freedom of Information Act request with the New York Police Department (NYPD) at the end of 2000. After several rounds of phone calls, I received my answer in spring 2002: my request was denied because (unlike Japan) the data do not exist (Reference GonzalezGonzalez 2002).Footnote ⁶

I also tried Los Angeles. With its low reliance on public transportation, Los Angeles is a difficult comparison to Japan, but the police function could conceivably be the same for lost items found on public property. The Los Angeles Police Department Property Division informed me that it logs all lost property. But it does not keep related statistics and in fact does not distinguish between property found by citizens and property retrieved by officers in the course of a crime investigation (Reference GoodrumGoodrum 2000).

I then turned to an alternative: Grand Central Terminal in New York City. According to a spokesperson for the regional commuter rail system Metro-North, the lost-and-found at Grand Central Terminal, which averages about 140,000 passengers per day, receives about 1,000 found items per month (Reference BruckerBrucker 2000). In 1995, the return rate was said to be 20% or less “at best.” Since that time, thanks at least in part to an enterprising former police officer who attempted to make the office more efficient through better labeling, the return rate for found items in 2000 was said to be “sixty percent or higher” (id.; Reference LombardiLombardi 1996).

The Grand Central figures are not scientific, and the fact that they were formulated by a Metro-North spokesperson may make them suspect. If they are accurate, they are high, approximately equal to the Japanese national rates. But Grand Central is a relatively closed environment, and as such may not be compared easily with the national Japanese figures.

For better comparison, I attempted to obtain figures from Japanese train stations. Although I was unable to obtain official figures from major Japanese railways, I was told unofficially that at Shinjuku Station in Tokyo, at which the average day sees 3 million passengers, according to East Japan Railway (and unlike Grand Central, it is a station on many lines, and not simply a terminal), the owner recovers about 80% of lost items. If that figure is accurate, it is significantly higher than that of Grand Central, despite much higher passenger volume.Footnote ⁷

In short, Japan generally appears to do a much better job of gathering data relevant to lost-and-found property than does the United States. The evidence on the efficacy of the system is equivocal, but the available data, combined with the lack of organized U.S. data and anecdotal evidence, suggest that the Japanese lost property system functions well in comparison with the U.S. system.

An Experiment

The absence of unequivocal comparative data calls for further investigation. To do so, I conducted an experiment based largely on various “lost” techniques of psychologists to study helping behavior, honesty, cooperation, and a variety of other social and political phenomena. Perhaps the most famous “lost” procedure is the “lost-letter technique” described and developed by Milgram (Reference Milgram, Mann and HarterMilgram, Mann, & Harter 1965). To avoid the pitfalls inherent in the survey process, Milgram devised a method of direct experimental observation:

In Milgram's first study, conducted in 1963, he found that while more than 70% of finders returned letters addressed to “Medical Research Associates” or to an individual, only 25% returned letters addressed to Friends of the Communist Party or Friends of the Nazi Party (Reference MilgramMilgram 1977:296–300). Subsequent studies, as well as modifications such as the “lost e-mail method,” found similar results (Reference Stern and FarberStern & Farber 1997).Footnote ⁸

Expanding on Milgram's technique, other researchers studied human behavior through “lost dollar” and “lost wallet” experiments. In “lost dollar” experiments, psychologists examined reactions to a lost dollar in a wallet, in an envelope belonging to an institution, and without an identifiable owner. Collapsing various situations and owner characteristics, they found that 31.2% returned the dollar, 39.6% ignored it, and 29.2% took it (Reference Penner, Summers, Brookmire and DertkePenner et al. 1976). Similar popular studies replicated these sorts of tests, finding, for instance, that 70% of wallets left on the street were taken, despite the fact that the wallets contained the owner's address (Reference Penner, Summers, Brookmire and DertkePenner et al. 1976:294). These results surely are discouraging to losers. Even when surveyed, a situation in which respondents can easily lie, 21% of 18-to-34-year-olds told Money magazine pollsters (Reference TopolnickiTopolnicki 1994) that they would keep a wallet with complete identification that contained $1,000.

Finally, in the most recent manifestation of the genre, economists Reference Dufwenberg and GneezyMartin Dufwenberg and Uri Gneezy (2000) conducted an “experimental lost wallet game” in which they attempted to use a bargaining game to determine the factors that might lead a finder to return a lost wallet. Not surprisingly, they found that the higher the stakes, the more likely the participants in the experiment were to keep the wallet. They also found some interesting altruistic effects, consistent with the Dictator game, in which one player decides how to divide a pot of money. But this experiment—and the discussion—rest on the assumption that players in the game, as well as wallet-finders in real life, have no economic reasons to report the find. The finder “simply keeps the wallet,” they note, when he or she realizes that “the owner does not reimburse [emphasis added] the finder if she picks her wallet up at the police station” (id.:64).

In Japan, as this article has shown, the dynamics are different. Although the owner does not fully reimburse the finder, leaving some moral hazard, the finders' fee and the possibility of full reimbursement after six months and two weeks encourage many people to turn in wallets.

To find out more about how these incentives operate in the real world and in cross-cultural settings, I conducted a similar test. In two locations in New York and one in Tokyo, I dropped objects and cash. New York's lost-property statute is relatively simple; objects with a value of more than $20 are required to be turned in to police (New York Personal Property Law § 252 2003). The statutory waiting period in New York for a find of less than $100 is 3 months, and the finder has a right to the property for 10 days after the expiration of that period (New York Personal Property Law § 253 2003).Footnote ⁹ Before the end of the period, the property goes to the loser if claimed.

My basic methodology (details are in the Appendix) was as follows. For objects, I dropped 100 mobile phones in each location. For cash, I dropped 20 wallets in each location. I obtained 60 wallets from a lost-and-found auction and inserted in each a small amount of cash and an identification card. In Japan, each wallet contained 2,000-yen bills. In the United States, each wallet contained two $10 bills to reach the New York $20 statutory minimum. This methodology may tell us little about how people react to valuable finds (such as a briefcase stuffed with cash), but it should also provide information about returns of more everyday finds.

I expected to find higher return rates for property than for cash. Property is relatively illiquid; an unknown person's mobile phone not connected to a service is not all that useful or valuable. Return of mobile phones may give a good indication of altruism, while return of cash may add an additional element of honesty. Of course, the actual calculation may differ by individual.

I dropped the phones and wallets in three locations. The first two locations were mixed business-shopping districts of Tokyo (Shinjuku) and New York (Midtown Manhattan). I designed the third location to test possible cross-cultural differences: I dropped objects in front of a New York grocery store (with the owner's permission) that caters almost exclusively to a Japanese expatriate clientele. I expected relatively good return rates from this location because of the clear identity of the person to whom the find should be reported. Still, if differences in lost-property recovery rates are based primarily on cultural traits, we might expect to see similar return rates among Japanese finders in Tokyo and New York. But if institutions are a large part of the difference, we might expect return rates to differ by country, not ethnicity.Footnote ¹⁰

The results of the drops were as reported in Table 2. As the table shows, for both phones and cash, the highest return rates came from Tokyo (95 phones, 17 wallets), and the lowest from New York (77 phones, 8 wallets, including two empty wallets).Footnote ¹¹ The differences are statistically significant (phones, χ ²(2)=13.1569, p<0.01; wallets, χ ²(2)=12.48, p<0.01). The same distribution was true for the ratio of objects returned via the police; 88 phones and 16 wallets were given to the police in Tokyo even though the objects contained identifying information that would have allowed that finder to contact me directly. In New York, almost all returns were made by phone calls to the phone mailbox listed on the object even though a police station was located nearby.

Table 2. Results of Property Drops

Finders' recovery rates tell a similar story. New York has no procedure by which a finder waives rights to claim unclaimed property, but in Tokyo, the finders of 83 of 88 phones waived such rights (wisely, as they had little value), while the finders of only 1 of 16 wallets waived (wisely, as the contents were worth 2,000 yen each). Of the five phones to which rights were not waived, two were recovered by finders after the waiting period, and of the 15 returned wallets to which rights were not waived, six were recovered by finders.Footnote ¹² In New York, only one of the wallets delivered to police was recovered by the finder after the statutory period (to the surprise of the police).

The results of the New York Japanese drop were interesting; return rates were higher than New York, but lower than Tokyo. Several explanations seem plausible. First, perhaps the drop location made returns easier than in the general New York case, as the most logical return location was the store clerk. But note that a few finders returned items to the police nonetheless, a response that does not comply with the law even in the Japanese system (which mandates return to the store), but may be an understandable product of a socially ingrained legal system, discussed in Part III, that encourages returns to police. Second, perhaps all Japanese finders are simply more honest or altruistic than non-Japanese counterparts. While this explanation is certainly possible, note that while the New York Japanese return rates were higher than the New York rates, they were lower than the Tokyo rates, suggesting that Japanese altruism is not monolithic. It may be that Japanese who come to New York are somehow “contaminated” by New York, a finding that could be explained by many different factors, including both formal and informal institutions. I return to these causal questions in Part IV.

Also worth noting is the fact that while no location had return rates as high as Tokyo, no location had abysmal return rates. Given that New York devotes virtually no public resources to the recovery of lost property, the return rates there may not be all that bad, suggesting, perhaps, that Japan is spending considerable resources for what may be a relatively small direct payoff. While Japan may be in fact gaining indirect payoffs in the form of greater societal comfort (which may result in greater economic activity), it may have the luxury of devoting its resources to what seems like a relatively insignificant endeavor because of its low violent crime rates. However, if police attention in Tokyo to such minutiae as lost bicycles is driving low rates of violent crime there, policymakers in New York and elsewhere would do well to take note. The available data do not unequivocally support either causal story, and some combination of both may be at work.Footnote ¹³

The Modern Japanese System

Japan's modern legal system has its origins in the late nineteenth-century Meiji Era (1867–1912). The Meiji legal reforms, based largely on French and German models, included the introduction of formal legal education organs and higher courts and the establishment of the first Constitution, Criminal Code, Code of Criminal Procedure, and Civil Code. Article 240 of the Civil Code, adopted in 1896 and still in force, provides that lost articles are the property of the finder if the actual owner does not appear within six months of the date of public notice of the loss, computed as two weeks after the recovery. The code then provides that the provision is to be administered “in accordance with other special laws” (Civil Code [Minpō] 1896, 1898).

The relevant “special law,” the Law Concerning Lost Articles (Ishitsubutsuhō 1899), was adopted in 1899 and remains in force. The Law Concerning Lost Articles provides a concrete set of rules from which to administer lost property. A person who finds lost property must return it to its owner or submit it to the chief of police within seven days of the find (Articles 1, 9). Lost property includes articles left behind by other people and domestic animals that have run away (Article 12). If a person finds lost property inside a private establishment (such as a department store, a ship, or inside the turnstiles of a railway), he or she must submit it to the management of the establishment within 24 hours (Article 10).

Importantly, the system applies to all property with no minimum value threshold. According to the law, a penny found on the ground cannot be pocketed but must be taken to the police.Footnote ¹⁴

The law establishes a reward system. Upon recovery, an owner “shall pay” the finder a sum of between 5 and 20% of the value of the lost property (Article 4). A finder has a civil right to the reward (assuming the find was reported within the seven-day period), but nonpayment is not a criminal infraction. If the property is found in a private establishment, one-half of the reward goes to the establishment owner, giving the establishment owner incentive to secure lost property (and less incentive for the individual finder in such establishments). If no one claims the property, and the finder waives his or her rights to it (or forfeits by [1] not turning in within seven days, or [2] being convicted in the past of embezzlement of lost property), then it becomes the property of the state (Articles 9, 15).Footnote ¹⁵

The carrots here, possession after six months and the finders' fee, can be significant. Japanese courts, normally not known for judicial activism, have attached some guidelines to aid in its administration. The value of the property is to be determined at the time of its return to the owner (Sugisaki v. Hayakawa Building Brokers Bank 1928). The exact fee within the 5–20% range depends on the type of property and the surrounding facts, but if the parties cannot agree, the court may determine the appropriate fee. The court generally compromises at 10% of the value (Nagaoka v. Okuyama 1922; Satō v. Shimizu K.K. 1965; see also Reference WagamatsuWagamatsu 1983:302; Reference ShimaShima 1986). Some special property deserves special treatment; negotiable notes are valued at one-third to one-half of face value, but non-negotiable notes are valued at only 2% of face value (Kōno v. Tokyo Renga 1991; Yoneda v. Tōkai Bank 1982; see also Reference KikuchiKikuchi 1998; cf. Ghen v. Rich 1881).

The carrots coexist with a potential stick. Article 254 of the modern Japanese Criminal Code, adopted in 1907 and still in force today, creates the crime of embezzlement of a lost article: a person may not wrongfully appropriate a lost article (Keihō [Criminal Code] 1907). In perhaps the best-known case in Japan, a group of fishermen in northern Akita prefecture, with the governor's permission, stocked a pond with carp. These were not just ordinary carp, but nishikigoi, big colorful carp that sell for $1,000 or so and eat out of your hand at Japanese gardens. The fishermen stocked their lake with about 1,000 of them. About 60 of them escaped their net and were subsequently “found” by Ono. Ono knew what he'd gotten, and he knew their source, but he nevertheless sold them for 380 yen per pound, for a total of 20,520 yen (they were supposedly worth 54,000 yen at the time). The court had little problem finding embezzlement and imposed on Ono a suspended sentence of six months' imprisonment (State v. Ono 1981).Footnote ¹⁶

Drafters of the Criminal Code apparently imported the criminal law category of “embezzlement of lost articles” from the German system, which made a distinction, subsequently adopted in Japan, between two types of embezzlement: lost-property embezzlement and entrusted-property embezzlement (Reference Satakura and DandōSatakura 1964:529). Entrusted-party embezzlement, as in the United States, is the misappropriation of property in custody. Unlike with lost-property embezzlement, the property is entrusted to the “finder.” In lost-property embezzlement, there is no such entrustment. In 1999, more than 98% of all embezzlement arrests in Japan (about 70,000) were for lost-property embezzlement.

A third category of property theft, larceny, is more difficult to distinguish from lost-property embezzlement. Larceny occurs when one deprives another of possession of an object; embezzlement occurs when one wrongfully appropriates an object that belongs to another. The key in distinguishing the two is possession (Reference TakahashiTakahashi 1987; see also Reference RoseRose 1985). Japanese courts have held that a camera left on a tour bus for five minutes within 20 meters of the owner remains in the owner's possession, so taking it would be larceny, not embezzlement (State v. Hosoda 1957). Bicycles are especially problematic and do not automatically invoke a theft charge. A bicycle left out at night less than two meters from one's home is still the owner's possession, and thus the charge is theft (State v. Yamamoto 1955), but if a drunken owner misplaces his bicycle, it is merely lost, so a subsequent taker is guilty only of embezzlement (State v. Masuda 1961).Footnote ¹⁷ Found objects, mail mistakenly delivered, and personal effects left on a train have all been held to constitute lost objects (State v. Sekiguchi 1948; State v. Gochi 1917; State v. Boku 1950).Footnote ¹⁸

In short, finders of lost property in Japan have three choices. First, they can ignore it with no consequence. Second, finders can turn in the property to the police or a private substitute. If they do so within seven days, they are entitled to either (1) the property, after six months and fourteen days, or (2) if recovered, a finders' fee of 5 to 20%. Third, finders may keep the property, but if they do so, they may be punished by fine or imprisonment.

History

The orderly scheme outlined above was adopted as part of the late Meiji reforms, in which German influence was strong (Reference HaleyHaley 1991:72). Some facial similarity to the German scheme exists. But closer analysis reveals a different, earlier origin, for as it turns out, these Meiji schemes were not novel, either. A nearly identical regime—never discussed in the secondary literature even in Japanese—functioned in the Edo Period, Japan's feudal era (1603–1868). In 1742, the shogunate promulgated the Reference Osadamegaki and OkunoKujikata Osadamegaki (1742) (roughly, “Official Collection of Legal Rules”), Japan's first attempt at a national compilation of pre-existing local laws for shogunate-controlled territories. Although the portion of the Osadamegaki (a book commonly called the Osadamegaki Hyakkajō, or Hundred Regulations) that contains the provisions pertaining to lost property was not for public consumption, the rules on which it was based were longstanding (Reference Osadamegaki and OkunoKujikata Osadamegaki 1742).

Article 60 of the Osadamegaki contains three provisions pertaining to lost property. First, the Osadamegaki incorporates a 1721 edict requiring that a finder submit lost property to authorities within three days. If the owner appears and the property is cash, he or she must halve it with the finder. If other than cash, the owner must pay the finder an unspecified reward. Second, incorporating a 1738 edict, the Osadamegaki states that if the owner does not claim the property within six months, possession is awarded to the finder. Finally, incorporating a customary rule of law with no fixed date, the Osadamegaki provides that a finder who appropriates the property will be fined.Footnote ¹⁹

While it is difficult to find the exact origin of the customary rule that fined those who misappropriate lost property, I was able to find criminal prosecutions from the late seventeenth and early eighteenth century that reflect the rule. Consider Kuemon's case, decided on July 6, 1728. Kuemon's daughter found a sword sheath on the road. She failed to report it. In fact, she sold it. Kuemon was fined three kan (very roughly $500–$1,000 today) (Tokugawa Kinrei Kō Goshū 1931). The earliest case I was able to find on point was Rizaemon's case of November 22, 1683. Although the details are unclear, Rizaemon apparently appropriated 50 ryō (very roughly $50,000 today) that his employer had lost, and hid it. Unfortunately for Rizaemon, his sister found it and reported the find to the authorities. It is unclear how the employer initially lost possession of the money, but it had occurred in such a way that the prosecuted crime was not one of larceny but one of embezzlement of lost property. The shogunate, perhaps noting the close relation to larceny (which was heavily punished) and the large amount of the loss, sentenced Rizaemon to death (Reference Harigamirui and JingōshichōHarigamirui 1896–1914).

After the end of the Edo Period, officials struggled to establish longstanding civil and criminal codes. The Edo Period criminal system was replaced in 1871 with the Shinritsu Kōryō (“The Essence of the New Code”), Japan's first national criminal law. Importantly, the drafters of the Shinritsu Kōryō were scholars of Chinese law, not Western law (the lone Western law scholar on the drafting committee was removed early in the process), and the resulting document strongly reflects the influence of the Chinese Ming Code (Reference TezukaTezuka 1984). The Shinritsu Kōryō, along with its supplementary successor, the 1873 Kaitei Ritsurei (“Amended Statutes”), states that a found article must be delivered to the authorities. Property not claimed within 30 days goes to the finder, but if the owner appears, then owners and finders share the property equally.Footnote ²⁰ A finder who fails to submit the property to authorities will be punished for embezzlement, but the punishment for embezzling private property will be one grade less than the punishment for embezzling public property, resulting in fewer beatings with a heavy stick (Shinritsu Kōryō 1871).

The 1871 system was supplemented in 1876 by the 1876 Lost Article Disposition Order, enacted by the Council of State (dajōkan) (Naikaku Hōkyoku 1876). The 1876 order, enacted as Western, and particularly French, influence was beginning to be seen in the legal system,Footnote ²¹ purported to be merely an administrative ordinance detailing the rules of the Shinritsu Kōryō system (Reference KiyouraKiyoura 1899), but in fact the precise regulations may have altered the incentive structure for some finders. The new system required a finder to deliver the property to officials within five days (Article 2; or eight days for animals, according to Article 9). It further stated that the property becomes the possession of the finder if the true owner does not appear within one year, and requires a reward of 5–20% of the value of the article to be paid to the finder by the owner (Article 4).

One final modification occurred to bridge the Edo system with the modern, more Western-influenced system. In 1880, the Council of State passed the first law officially dubbed a “penal code.” The 1880 penal code (Kyūkeihō), which stood for 27 years before the government replaced it with the modern version, altered little of the definition of lost-property embezzlement but changed the penalty. Now, instead of blows from a heavy stick, one who embezzled lost property would receive imprisonment for 11 days to three months as well as a fine of two to twenty yen (Kyūkeihō 1880, Article 385). The reduction of the penalty, enacted as part of a broader copying of French criminal law, was part of a larger movement to relax the severity of punishment (Reference IshiiIshii 1992:43–52).

Table 3 details these various changes in the past three centuries in the lost property regime. In summary, although various foreign influences can be seen in presentation, penalties, and the like, the early Meiji schemes are virtually identical to the 1899 law that is still in force today, and the basic structure is similar to that of the Edo Period. Thus, depending on how one counts, the system has been in existence for at least 125 years and perhaps in excess of 300.

Table 3. Historical Development of Lost Property Law

Prosecution and Enforcement

Of course, few modern cases of embezzlement of lost property are punished by death. But cases are indeed reported, and some are indeed prosecuted.

Arrests

For nontrivial cases, the pipeline begins with arrests. Consider first the annual numbers of arrests and cases sent by police to prosecutors, as shown in Figure 5 (arrests and prosecutions from Reference HōmushōHōmushō (1974–1999). The figure indicates a rising and not insubstantial number of arrests each year. The number of cases sent to prosecutors post-arrest is approximately equal to the larceny rate and more than 20 times the murder rate.Footnote ²³

Figure 5. Incidents of Embezzlement of Lost Property, 1989–1999

Prosecution: Juveniles

Figure 5 reveals only the number of persons who meet thresholds for arrest and forwarding to prosecutors—prosecution itself is another matter. Consider next the prosecution figures. Figure 6 details the total number of cases and the two major categories of disposition for lost-property cases during the period 1973–1999 (Reference HōmushōHōmushō 1974–1999).

Figure 6. Prosecutorial Disposition of Lost-Property Cases, 1973–1999

The figure shows that more than three-quarters of all cases are sent by prosecutors to family court; in other words, more than three-fourths of the offenders in lost-property embezzlement crimes disposed of by prosecutors are minors under age 20. Teenagers are most likely to commit the crime; the distribution of juveniles disposed of by family courts in lost-property embezzlement cases in 1990–1999 is roughly 28% for juveniles under 16, 41% for juveniles ages 16–17, and 29% for juveniles ages 18–19 (Reference Saikō Saibansho, JimusōkyokuSaikō Saibansho Jimusōkyoku 1991–2000).

The distribution of crime among adults and minors is significant and implies several possibilities. Perhaps minors are treated more leniently than adults: adults might be charged with larceny (with penalties of up to 10 years), while juveniles are merely charged with lost-property embezzlement (with penalties of up to one year), which would distort the statistics. While this explanation is certainly plausible, arrest rates tentatively suggest that it may be incomplete. Although adults do indeed comprise less than 25% of the lost-property embezzlement pool, they only account for 43% of the larceny pool.Footnote ²⁴ Moreover, larceny is the only crime for which juveniles are arrested more than lost-property embezzlement. Larceny accounted for 65.8% of all juvenile arrests in 1998 (Reference HōmushōHōmushō 1999), suggesting that larceny remains an option for juveniles as well.

My interviews with police and offenders suggest a different explanation—juveniles actually do commit more lost-property embezzlement than adults. The 1999 White Paper on Crime (Reference HōmushōHōmushō 1999) states that more than half of lost-property embezzlement arrests of juveniles are over misappropriated bicycles. The most common scenario, according to both police and suspect interviews, is the taking of bicycles from train stations. When the ownership of the bicycle is clear, larceny is generally charged; if not, lost-property embezzlement is charged, suggesting that these arrests are for misdeeds more akin to theft than to the everyday lost-property situation. But bicycle cases are not the object of misappropriation. Police, prosecutors, and juveniles to whom I spoke detailed a long list of embezzlement scenarios not listed in any statistical survey: a cornucopia of failures to return or report cash, wallets, shopping bags full of purchases, mobile phones, purses, electronic games, jewelry, skis and other sporting equipment, and so on. Although I have no systematic method of matching loss reports with state enforcement, police whom I interviewed estimated roughly that about half to three-fourths of lost-property embezzlements could be so linked.

A juvenile sent to family court for any crime faces one of five dispositions: no trial (the equivalent of dropping charges for an adult), no decision (equivalent to acquittal), counseling, a decision to protect through probation or institutional reform (equivalent to guilty), and referral back to the prosecutor's office based on the seriousness of the crime or the fact that the offender has legally become an adult subsequent to the arrest (Reference GotōGotō 1997; Shonenhō [Juvenile Law] 1992). The last option, potentially the most serious, is used very sparingly in lost-property cases. Of the annual average 27,497 cases decided in the 10-year period beginning in 1989, only about 17 annually were sent back to prosecutors, and fewer than two annually because of their “seriousness.” Counseling is also rarely used; only five annually fell in this category. The primary disposition, used in more than 90% of the cases, is no trial, the equivalent of dropping charges. About 4% receive no decision, the equivalent of not guilty. About 1% are found guilty, and of those, an annual average of about 232 are given probation, five are sent to reform school, and 21 are sent to a juvenile correctional institution.

Comparison to the United States

Compared to the Japanese system, the U.S. lost-property legal regime is complex and unpredictable.

Evidence on Honesty and Altruism

All other things being equal, we might expect honest, fair, and altruistic people to report lost property more often than their opposites. If Japanese people are more honest and/or altruistic (two separate concepts that I discuss together here), we would expect more recovery of lost property in Japan. But the available evidence—and it is sparse—from sociology, psychology, economics, and anthropology is ambiguous at best. Consider the following four studies, the first three of which are comparative:

• • From sociology, in his comparative study of trust, Reference YamagishiYamagishi (1998) surveyed 928 university students and nearly 500 randomly sampled persons in Japan and the United States. Yamagishi asked several questions directly related to honesty. In response to the statement “I keep in mind the spirit of fair play in every situation,” Yamagishi found statistically significant higher levels of honesty in American men in both the university and random samples, and no statistically significant differences among American and Japanese women. In response to the more direct statement “I do not want to be dishonest in any situation,” Americans rated significantly higher in virtually every subsample. From these questions and others, Yamagishi constructed an index of honesty and fairness in which Americans rated significantly higher than Japanese in every category.Footnote ³³

• • Psychological studies (Reference Kawakami and Takai-KawakamiKawakami & Takai-Kawakami 1995) examine honesty and helping behavior in children. Two Japanese researchers attempted to measure deception in Japanese and American 3- to 6-year-olds by hiding a toy and telling a child not to peek. Most children peek, and most children lie about peeking. But the experiments revealed no significant differences among U.S. and Japanese children in regard to deception.

• • In an economic study of altruism, Horioka and three co-authors (Reference Horioka, Fujisaki, Wanatabe and KounoHorioka et al. 2000) analyzed the “Comparative Survey of Savings in Japan and the United States,” a survey conducted in 1996 by the Japanese government. Using a variety of data on saving and bequest motives in the two countries, they found the evidence to be “remarkably consistent” with the theory that “the selfish life cycle model,” in which “individuals are selfish and do not care about their children or about anyone else,” is “the dominant model of household behavior in both countries but that it is far more applicable in Japan than it is in the U.S.” (Reference Horioka, Fujisaki, Wanatabe and KounoHorioka et al. 2000:2).Footnote ³⁴

• • From anthropology, in another study on helping behavior in adults and children, Reference LebraLebra (1976) administered a thematic apperception test in which she asked approximately 200 Japanese adults and high school seniors to finish the phrase “If you are kind to others…” The two most popular categories of responses were “autistic satisfaction” such as “You will feel better” or “Your feelings will be enriched” (34.3% for adults, 34.1% for high-schoolers) and “reciprocal return in other forms” such as “You will be rewarded” (30.3% for adults, 31.9% for high-schoolers). But among male high-school students, the reward was the most popular answer (35.6%).

I do not mean to suggest here that Japan somehow revels in dishonesty, or even that altruistic behavior is any less prevalent in Japan than elsewhere. Nor do I claim that the evidence on honesty, altruism, and social norms constitutes the entire spectrum of Japanese society and culture. Cultural differences may lie elsewhere, such as in concepts of luck and desert, underlying class structure and resource distribution, views regarding the value or nature of ownership,Footnote ³⁵ or the actor's concept of self and individual responsibility (Reference Hamilton and SandersHamilton & Sanders 1992), and this article does not attempt to investigate the entirety of culture.Footnote ³⁶ I merely note here that after searching the literature for data that might shed light on the ethical model that may be most relevant to lost property in Japan, the above studies seem to constitute some of the best available empirical evidence on the topic, and their results are far from certain.

The studies suggest a conclusion and a conjecture. As for the former, at least in comparison to the United States, the Japanese propensity for returning lost property does not appear to be due largely to more widespread or deeply rooted Japanese norms of honest behavior. Some other factor or set of factors is likely at work.

As for the latter, this article has shown a striking correlation between lost-property practice and lost-property institutions in Japan. The correlation can lend itself to two opposing interpretations of historical causality: either lost-property practice spawned institutions, or lost-property institutions encouraged the development of lost-property practice. Although the historical record can support either interpretation, there is an analytical way to gain a better—though very tentative, I recognize—understanding of the causal relation. Assume that honesty can be divided into two types, “general” and “specific.” Specific honesty is the practice of honest reporting of lost property; general honesty is the broader phenomenon examined in the studies above. General honesty, as some of the studies above suggest, is arguably similar in Japan and the United States, but lost-property institutions in those two countries differ. Assuming that specific honesty and general honesty are correlated, it seems relatively unlikely that differences in the practice of specific honesty have led to institutional differences (in the form of codification of practice), as similar general honesty structures should have led to similar institutional structures in both countries. The fact that institutional differences have arisen despite similar honesty structures leads to the conjecture that Japanese institutions may have “carved out” specific honesty as separate from general honesty.

Still, general honesty may not necessarily correlate with lost-property specific honesty, and applying these studies directly to the lost-property context may be problematic. Although honesty and fairness may play a role in some decisions of whether to report lost property, other concerns (including not only the potential cultural factors listed above but also the business of one's schedule, for instance) may be important as well. It is also unclear whether these studies of largely interpersonal honesty measure the same sort of factors that constitute the sort of societal helping behavior that characterizes lost-and-found. Accordingly, I conduct two new studies to more directly examine the lost-and-found context.

Survey

In 1939, Reference DonnerDonner (1940) surveyed 2,188 elementary and secondary school students in Iowa and Texas regarding their knowledge and opinions regarding finders' law and their reactions to various lost-property hypotheticals. She found, first, that 86% of the students' judgments on the hypotheticals were roughly in line with existing case law. But she also found:

Donner's study raises many questions. What does “give the state statute,” for instance, mean? Despite the survey's eccentricities, it is interesting as a serious effort to examine the phenomenon, and Donner's sample choice, at least, is intriguing. Note that one of the jurisdictions tested, and in fact the source of two-thirds of the students in the sample, was Iowa. As discussed above, Iowa has since 1839 had a statutory scheme that entitles a finder to 10% of the value of his or her find. The timing is also interesting; just five years before Donner's survey, in a case that accounts for 11 of the 13 references in the Iowa Code Annotated, the Iowa Supreme Court (Flood v. City Nat. Bank of Clinton 1935) decided the constitutionality of the provision and declared that it served laudable public policy goals. If there has been any U.S. jurisdiction at any point in time when knowing the law could be particularly advantageous, or perhaps when we might expect people to know the law, it might have been Iowa in 1939. Yet none of the students in Donner's survey knew the law. I would not expect the results of a more recent survey to differ considerably.

For comparison, with the help of assistants, I conducted a simple survey regarding lost property of 615 ninth-graders in Japan. The written survey contained four open-ended scenario-type questions. The instructions stated that the test was one of legal knowledge. The primary questions and summarized responses were as follows:

1. 1. A finds a wallet on the sidewalk with 30,000 yen inside, and takes it to the police. B, who lost the wallet the previous day, goes to the police and claims the wallet. The police give B the wallet. Does A have a legal right to collect a reward from B? If so, how much? 497 students (80%) correctly responded that A has a legal right to a reward. An additional 30 students incorrectly answered that A has no legal right to the reward, but still should receive one. Of the 497 who answered correctly, 410 responded with award amounts that fell in the correct range of 5 to 20% of 30,000 yen. Most (339) stated that the reward would be 10% or 3,000 yen.

2. 2. A takes the wallet in Question 1 to the police. The person who lost the wallet never recovers it. What happens to the wallet and the money inside? In one formulation or another, 510 (83%) students correctly stated that the finder has a right to the property. The most common other answer was that the police take the property.

3. 3. After finding the wallet in Question 1, A keeps it. Is this a crime? Five hundred fifty-two students, or about 90%, correctly stated that keeping the wallet is a crime.

4. 4. A sees the wallet in Question 1 and in fact sees money sticking out of the top, but he ignores it and keeps walking. Is this a crime? Three hundred thirty-two students, or 54%, correctly stated that ignoring the wallet is not a crime.

The survey data suggest two primary points. First, Japanese students seem to have a high degree of knowledge of lost-property law; or, alternatively, lost-property law might coincide with norms,Footnote ³⁷ and students have knowledge of those norms. The origins of this knowledge appear to be dual; home, as discussed above through kōban visits, and school. As for school, lost-property law per se is not part of the Japanese classroom curriculum; teachers clearly are not reading the statute in class. But the Japanese elementary education system's holistic focus “on children's long-term internalization of values” (Reference LewisLewis 1995:212) may lead to tacit knowledge of the rules, or perhaps to an inculcation of the norms that underlie lost-property institutions, regardless of whether students can “give the statute.”

A second point worth noting from the survey data is that when students erred, they generally erred on the side that would facilitate the return of lost items. In response to Question 4, nearly half of the students answered incorrectly that ignoring the wallet is a crime. Although the response is incorrect, if such opinions are widespread in the population, they might lead to a high level of recovered lost property.

Interviews

The data presented above and in Part I suggest that (1) Japanese people know finders' law, and (2) measurable differences in lost-property recovery can be seen among New Yorkers, Tokyoites, and Japanese expatriates in New York. But the data do not necessarily show a linkage between (1) and (2); that is, that law accounts for the differences. Japanese lost-property recovery rates might be higher not because of the law, legal institutions, or enforcement, but because of social norms or altruism. The conjecture raised earlier in this part—that institutions may have spawned practice—refers only to the origin and direction of historical causality (which came first, institutions or practice?), and not necessarily to the daily thought process that characterizes motivational causality (which plays a larger role in explaining actions, institutions or practice?). To explore these causal factors, I turn to interviews.

I was able to interview 38 persons from the Tokyo experiment, 22 persons from the New York experiment, and 10 persons from the New York Japanese experiment. Even without additional analysis, these numbers say something about the lost-and-found process. The number was relatively high for Tokyo, as finders are requested to fill out a found property report by the police that lists contact information. The number was perhaps higher than one might expect in New York, as most persons who returned the property contacted me by phone instead of using the police or a merchant as an intermediary. The number was relatively low for New York Japanese, as most finders simply returned property to the store clerk and left no additional information. Of course, in all cases, more interviews are better, but I simply was unable to interview any additional persons in the dataset.

Through these interviews, I attempted to ascertain why persons returned lost property. Three primary groups of answers were given in all three locations. The categories have some overlap, but some basic distinctions can nevertheless be made. First, some explanations focused explicitly on formal institution–based rewards. I recorded these answers, such as “If the owner doesn't claim it, I have a right to it” or “I want the reward money,” as institutional explanations. Importantly, for an answer to be coded as “institutional,” a returnee did not have to be able to recite the statute or know where to find it. A finder did not even need to be aware that the rule was a creature of the state, though I attempted to determine more precisely finders' knowledge on this point.

These institutional explanations are relatively easy to distinguish from the second and third categories of answers, which were more difficult to separate. The second category includes explanations that centered on social norms and civic duties that may be informally enforced but are unregulated by formal institutions. I recorded these answers, such as “It's what you're supposed to do” or “Everybody does that,” as social norms explanations.

Finally, some explanations were based on notions of altruism and similar internal other-regarding preferences that had no obvious link to either formally enforced institutions or informally enforced social norms. While some of these internal factors may in fact be internalized social norms (Reference McAdamsMcAdams 1997; Reference StoutStout 2001), I simplified the inquiry by categorizing answers such as “Honesty is the best policy” as other-regarding.

Cross-cultural surveys are of course imperfect; it is possible that observed differences in result are based on differences in the way that a question is understood or in interviewees' response styles. As for perception of the question, although I attempted to mitigate the problem by vetting the survey with experts and amateurs in both countries, differences may arise nonetheless.

As for response styles, in this particular case, social psychology literature (Reference Heine and RenshawHeine & Renshaw 2002; Reference Kitayama, Marcus, Matsumoto and NorasakkunkitKitayama et al. 1997) suggests that Japanese people have low self-esteem and are self-critical (for instance, they are reluctant to conclude that they perform better than average classmates), while North Americans have high self-esteem and are self-enhancing (and are reluctant to conclude that they perform worse). Given these studies, one might argue that Japanese interviewees would be less likely than New Yorkers to give self-enhancing explanations that boast of their honesty,Footnote ³⁸ which could result in fewer “other-regarding” responses. While no study can completely avoid such difficulties, in this case, the inclusion of relatively self-neutral responses in the other-regarding category may mitigate the concern. For instance, “I felt sorry for you” or “So sad!” (“Kawaiiso”) was a common response of Japanese interviewees and is an often-used Japanese phrase. While some scholars (Reference SallySally 2002:456–57) formally distinguish sympathy from “pure” altruism, following other empirical studies of altruism in Japan and the United States, (see Reference Matsui, Nakasato and IshiiMatsui, Nakasato, & Ishii 1998), I include them together in the third category of other-regarding responses. If, as Reference SallySally (2002) discusses, sympathy is based on social interaction, physical distance, and psychological affinity, it would not be unreasonable to expect a large percentage of Tokyo responses relative to New York responses to fall in the “other-regarding” category, but other interpretations of distance are of course possible.

Crucially, very few interviewees gave only one type of explanation. I found that if I listened carefully and persistently, interviewees tended to offer two, and sometimes all three, explanations. I recorded all explanations. Two primary factors motivated this strategy. First, by not forcing interviewees into a single explanation or a binary choice, I hoped to mitigate further some of the cross-cultural concerns of the previous paragraph, because multiple responses may allow for more self-praise or self-criticism, and because there is no a priori reason for a self-critical person to choose one self-neutral option over another. Second and more basic, human motivation is complex, and while multiple responses may be somewhat imprecise, they may be better able than binary choices, Likert scales, or short-answer surveys to capture some of this complexity. As Table 4 shows, the mix of responses differed enough to suggest real differences among the locations.

Table 4. Explanations for Returns of Lost Property

When asked why they returned the property, Tokyo finders typically offered one of two explanations. First, many saw returning lost property as a civic duty and explained that such values had been inculcated in them since childhood. About half of these social norms explanations also included mentions of kōban.Footnote ³⁹ While these explanations touched on institutions, and I regard kōban themselves as an institutional variable, because of the focus on civic duty and values, I recorded such responses as social norms explanations in the absence of more explicit formal institutional mentions. Had I included these explanations in the institutional category, that response would have accounted for more than 90% of the responses. In any event, kōban explanations were not honesty-based, and arguably were not altruistic or otherwise other-regarding.

Second, many Tokyo interviewees, often in addition to the social norms explanations, volunteered more explicit institutional carrot-and-stick explanations. Seven interviewees cited the stick, noting that not turning in the property was a punishable form of theft. All 38 Tokyo finders knew of the reward system, and 10 referred explicitly to the reward as “one-tenth” (ichiwari). To be sure, most finders did not know the exact fee mandated or other statutory details. But all knew of the system's existence, and only two finders seemed unsure as to whether the reward system was created by formal institutions (one thought it was a local police custom, another thought it to be merely urban legend).Footnote ⁴⁰

At least in the case of the phones, the carrots and sticks may not be explicit. A finder is always free to ignore the phone, eliminating the stick. Neither does the carrot seem large or even certain; 10% of the value of a phone often is a pittance, finders seldom need another's phone after six months, and a finder's only remedy against an owner who refuses to pay a reward is a civil suit—all plausible reasons why the waiver rate for phones in Tokyo was relatively high. Still, the Tokyo interview responses suggest that people act as if the carrots and sticks matter, whether because of incomplete familiarity with the law, because of the ingrained nature of the system, because the thought process does not begin until the object is already in one's possession, or for some other reason or combination of reasons.

Sympathy-based predictions notwithstanding, it was in New York, not Tokyo, that the reasons for returning lost property were more likely to be stated in other-regarding terms.Footnote ⁴¹ New York interviewees stressed that “I'm an honest person,”“I couldn't live with myself if I kept your money,” and “I just thought about how [inconvenienced] I'd be if I lost my phone.” I classified these explanations as “other-regarding,” and while I recognize that they might also be “social norms,” they certainly are not “institutional.” One vocal male interviewee put it more succinctly. After he gave me several individual-honesty explanations, I tried leading him by saying, “Oh come on. I'm sure a lot of people return stuff just because they want a reward, right?” He responded, “You brainiacs might think so, but you don't know human nature. Not everybody does everything for money. I called you because it's the right thing to do.” Only a handful of New York interviewees focused on social norms, and only one mentioned the possibility of reward—and he knew of no legal basis for it (none exists).

The results of the New York Japanese experiment interviewees, like those of the experiment itself, fell somewhere in between those of Tokyo and New York, though the small sample size yielded less-than-conclusive results. Interviewees focused largely on institutions but seemed to give greater weight to altruistic factors than did the Tokyo interviewees. Few spoke of social norms, perhaps because they did not feel familiar with New York norms. Interestingly, however, while these interviewees seemed confident in their institutional assessments, their knowledge of New York institutions often was simply incorrect. Many New York Japanese interviewees incorrectly assumed an institutional reward structure similar to that of Japan. Some were surprised to hear that New York has no kōban system (most New York Japanese made returns to the store clerk). In other words, New York Japanese relied heavily (80%) on institutions, but got the institutions wrong.

From this finder interview evidence, I conclude that formal institutions are a central, but certainly not the only, factor in the recovery of lost property. In Tokyo, where finders' law and related institutions are good and recovery rates are high, interviewees focused on institutions. Among New York Japanese, where recovery rates were good but not as good as Tokyo, interviewees continued to focus on institutions despite the fact that they got the institutions wrong. And in the general New York setting, in which finders' law institutions are poor and recovery rates are relatively low, the few New Yorkers who returned the property did so for reasons independent of institutions.

The New York setting in the comparative context further suggests that while altruism gets results, altruism alone may not provide the same benefits as altruism plus institutions. It may be that finders are in a moral “double-bind”; they are “guilty” if they keep the property and “suckers” if they turn it in. Institutions like those in Tokyo may be one factor that encourages people at the margin to solve the double-bind by reducing their “sucker” feelings through rewards. Of course, these marginal calculations are imprecise and muddled. While institutions may result in a net increase in returns, they may in fact do so via a reduction of underlying altruism as the reward incentives “crowd out” internal ones (Reference Frey and JegenFrey & Jegen 2000).Footnote ⁴²

To obtain an additional reference point, I supplemented these post-experiment interviews with police interviews in Japan. In the course of researching this article, I spoke to 40 police officers at thirty kōban. Each of these officers is engaged in lost-property “practice” of some sort on a daily basis. Again, officers gave multiple explanations. Like the post-experiment interviewees, most of these officers (37) suggested that people return lost property because of institutional concerns. Thirty-two suggested social norms, and twenty suggested altruistic explanations.

The data in this section warrant a final caution. Although I have categorized explanations for analytical clarity, I make no claim that law can be separated from underlying societal and other informal institutions. The relation between law and society is undeniably complex, and the lost-property context is no exception. On one hand, a primary reason for the Japanese lost-property regime's consistency over time may be that it reflects established informal practices that might function effectively today even in the absence of formal institutions.Footnote ⁴³ On the other hand, the general-honesty evidence on historical causality tentatively suggests that as the law has become more embedded in society (Reference Aoki, Aoki and HayamiAoki 2001; Reference GranovetterGranovetter 1985), it standardized nascent norms, or perhaps even created them from scratch, suggesting greater institutional importance. Then again, perhaps the relation is better described as one of mutual reinforcement, as suggested in the kōban discussion above.

The issue is complex, and I see little analytical value in imposing a total dichotomy here. But the data presented above may at least offer insights into individual perceptions and conceptualizations of their behavior on this issue at a static point in time. The individuals in my interviews appear to have had little problem categorizing “law” separately from other explanations. In Tokyo, everyone knew the rules, almost everyone knew those rules were formal, and most said that they relied on them. In New York, virtually no one knew the rules, and almost no one said that they relied on them.

With these cautions in mind, the evidence presented above thus suggests that finders are motivated by a complex interaction of altruistic and legal factors. The available evidence also suggests that conceptions of correct behavior are strongly intertwined with the legal environment into which one is socialized. Whether the specific lost-property institutions that comprise the Japanese legal environment are a product of such conceptions, a cause, or both, they play a central role in explaining the efficiency of the Japanese system.