I have been led by my efforts to understand China tosee social interaction in terms of contractual relation-ships.When I was in Guangzhou in 1979，I was struckby the meticulous detailing of rank in different occupa-tions.A certain rank would entitle a comrade to share anaut om obile，or to an egg every other day，or to the right to buy fish without having to stand in line.These phe-nomena were thought provoking.My first explanationwas that since people are born unequal，in a “property-less” state where everybody is equally “property-less”,human rights must be unequal in order to produce socialequilibrium.It took me two years to see the deepertruth: The hierarchical rankings adopted in China arecontractual restraints essential to reduce rent dissipationunder competition，in a situation where the delineation ofrights in terms of property is absent.
【14】Cheung，Will China Go Capitalist？（London 1982），HobartPaper 84，Section II.
【15】Price is a constraint which restrains competition.As AdamSmith puts it in Wealth of Nations: “Give me that which I want,and you shall have this which you want.”（Cannan edition，p.18）.A market price is implied.
The important implication here is that economicreformin China involves a shift froma system of rightsin terms of hierarchy to a systemof rights in terms ofproperty，or a change from one type of contractualarrangement for restraining competition to another type.This，I profess，is the true explanation of what has hap-pened these 30-odd years.The fact that the Chinese managed to succeed without social upheaval may beregarded as miraculous，and as we shall see，the key tosuccess is that they used a contract sitting in between,known as the responsibility contract.The true miracle is not only that they manage to do it，but the system they end up with.
While we are on the level of general theory，let mepoint out that restraining competition in resource use is costly，and these costs are somewhat misleadinglycalled transaction costs.I have emphasized over many years that different types of transaction costs can only be separated at the margin，so that testable implicationscan be obtained by specifying marginal changes in these costs.I have emphasized also that it is not essen-tial to measure transaction costs in（real）dollars andcents，because it is sufficient to measure in terms of ourability to rank transaction costs under different observ-able situations.No easy task，but it can be done，and I have obtained on numerous occasions predictions orexplanation based on specifying observable changes intransaction costs in this sense.You may quarrel withmy explanation of why better seats are underpriced,16but over the years the accuracy of my predictions ofevents unfolding in China have earned such high marks that they cannot just be the results of gazing at a crystalball.
The impossibility of separating different types oftransaction costs except at the margin has forced me tobroaden the concept to include all costs that do notexist in a Robinson Crusoe economy.As a result,transaction costs often exist in situations where market transactions do not exist.I prefer to use the terminsti-tution costs，or costs that exist only in a society.It is my contention that transaction/institution costs arisemainly from restraining competition in the use of scarce resources，or，in my broadened notion of con-tractual arrangements discussed earlier，from the requirement of using contracts to restrain competition.
【16】Steven N.S.Cheung，“Why Are Better Seats ‘Under-priced’？” Economic Inquiry（1997），pp.512-522.
The upshot is that so long as competition exists，trans-action/institution costs must also exist.In other words,to speak of a society without these costs entails a con-tradiction in terms.
In 1982，I said that if there were no transaction/institu-tion costs there would be no market.Commenting on the Coase Theorem，I wrote:
If all transaction costs，broadly defined，were trulyzero，it would have to be accepted that consumerpreferences would be revealed without cost.Auctioneers and monitors would provide free all the services of gathering and collating information;workers and other factors of production would bedirected freely to produce in perfect accord withconsumer preference; and each consumer would receive goods and services in conformity with his preferences.The total income received by each worker（consumer），as determined costlessly by anarbitrator，would equal his marginal productivityplus a share of the rents of all resources according toany of a number of criteria costlessly agreed upon.By such reasoning the Coase result can be obtainedwithout a market price.17
That the market exists because transaction/institutioncosts are not zero is a viewconsistent with Coase’s clas-sic analysis of the firm，and my early work on the choice of contractual arrangements.
【17】Cheung，Will China Go Capitalist？op.cit.，Section Ⅲ.
It follows almosttautologically that markets emerge to reduce transactioncosts.However，unlike explaining theater-ticket pricingor the choice of buffet-dinner arrangements，the inter-pretation of which requires only the specification of cer-tain marginal changes in transaction/institution costs，tointerpret complex and involved economic systems orchanges in such systems entails a far higher level of dif-ficulty.
My difficulty，which lasted for nearly 20 years，was the result of a mental block.I did not know what kind oftransaction/institution costs should be introduced tointerpret the existence of private property rights and the market.The broad viewI have assumed indicates that these costs are everywhere，and there is no roomto addanymore.Then，one night in 2001，I sawthe light: We have to deduct，not to add，these costs to obtain the solu-tion.