166the oxford introductions to u.s. law: property
who wants to know who the current owner is. So when A grants to B, A must transfer a deed to B’s control. Borderline cases become interesting, as where A puts a deed in a place where B will find it after A’s death, as a substitute for a will devising the land to B. If the location is under B’s control, and A cannot easily get the deed back, it is probably a delivery. Conditional deliveries can receive very different treatment in different jurisdictions because two policies come into conflict. As we saw with gifts causa mortis (see Chapter 4), courts do not look with favor on parties’ attempts to evade the formal requirements for wills, which give an assurance of the testator’s intent and a lack of coercion, and so on. However, it is well known that probate is cumbersome, and the well-advised can avoid it by using trusts. So A’s delivery of a deed to a trusted person with conditional instructions for delivery (e.g., after the grantor’s death) might be regarded as a “poor man’s trust,” and if so, giving effect to the deed makes sense.
• Title Records
The most elaborate notice-giving devices of modern times are the systems of title records. The most widespread title records are for property in land. Because land is fixed in location, valuable, and used for secured lending, it has lent itself to systematic record keeping, in America since Colonial times. Some types of personal property such as cars, boats, airplanes, lost artwork, and some intellectual property also have registries. But most personal property is not registered except in connection with its use as collateral for security interests (see below).
Land records in the United States are highly decentralized. Although the rectangular survey is national, and titles to much private land trace back to a grant from the federal government, title records are typically kept at the county level. Most land in the United States is covered by recording systems or recordation, in which records relating to title (deeds, mortgages, etc.) are filed in an
land transactions and title records 167
office, and indexed by grantor and grantee. (A few jurisdictions also have tract indexes in which all the documents relating to a given parcel are referenced.) The other major type of system is registration, in which land records are typically arranged by parcel, and the land office inspects each document to be filed and guarantees the validity of titles. The Torrens system and the systems of many civil law countries exemplify registration, which has a higher start-up cost and requires better trained personnel (the more so the more definitive the title documents are meant to be).
Registration provides a more definitive answer to ownership questions than does recordation. Registration avoids more errors, and those errors that do occur are insured through registrar liability with compensation funded through registration fees. In recording systems, this insurance function is provided by private companies, which typically maintain their own version of the title records (called “title plants”). Title insurance provides a level of assurance that probably is sufficiently effective to make a shift from recordation to title registration not worthwhile once a system of recordation is established. However, countries setting up land records for the first time, as in former communist countries, tend to opt for registration, whether out of its advantages in terms of accuracy or its association with civil law, especially of the German type.
In a recording system, there is no certification of title by public authorities. The only function of the public records office is as a repository of records, organized by indexes. It is the responsibility of prospective buyers and lenders (and their attorneys) to examine these records and figure out for themselves whether the seller is offering a sufficiently good title. The central concepts in conducting these examinations are the title search and the chain of title. A title search involves going backward in the grantee index to find the links in the chain of title (for example, C received from B who received from A). This is less than half the story, however. The real worry is that any one of the people in the sequence might have done something untoward such as convey the same land twice. So one must go back down the grantor index to find out
168the oxford introductions to u.s. law: property
what these people have done. The most thorough conceivable search would involve figuring out what each person in the chain of title might have done at any time. This would be costly and largely pointless. Instead, buyers are only responsible for knowing what each link in the chain did from the time that link received a deed until the time the next person recorded. Why is that? Because in certain circumstances, persons who fail to record may lose their interest in the property to persons who acquire interests later in time.
How is it possible for C to receive good title from A when A had already conveyed the land away to B? In a world of no title records, the answer would be that C could not. The foundational principle of property transfer is nemo dat quod non habet (“one cannot give what one does not have”), and as we saw in Chapter 5, a grantor is presumed to give all that the grantor has unless specified otherwise. That is, the transferee steps into the transferor’s shoes except to the extent that a lesser interest is being carved out for transfer. If the shoes are already on some other transferee’s feet (metaphorically speaking in all cases except for shoes), then A has nothing left to transfer. The effect is like first-in-time. If A transfers to B and then to C, there was something to transfer to B the first time but not to C the second. (If A purports to transfer to B property A does not have and then later acquires it, B can sue to force a transfer to B under the equitable doctrine of estoppel by deed.)
Recording acts are superimposed on the basic nemo dat regime. Thus, if a recording act does not apply, nemo dat does. The main effect of a recording act is to cause someone—typically a later transferee—to win who would otherwise lose under nemo dat. There are three types of recording acts. The earliest type, which is still found in North Carolina, is the race statute, under which the first claimant to record wins. Thus if A sells to B, and then A sells to C, but C records before B, C has better title than B. (If B had recorded immediately, the recording act would not change the result from nemo dat, that B beats C.) What can B do? B can sue A, if A can be found and is not judgment-proof.
land transactions and title records 169
Courts were none too pleased when people in the situation of C, the later transferee, took the deed knowing of the earlier conveyance. For C to win under such circumstances seemed unfair, and there could be many reasons why losing the property might seem a disproportionate punishment for B’s lack of timely recording. So courts began to make exceptions, which culminated in notice statutes under which nemo dat is only overridden in favor of subsequent good faith purchasers for value (GFPVs). A purchaser for value is someone who gives consideration for the property. Thus, a person taking by gift or inheritance would not be able to take advantage of the statute. Someone is in good faith if they have no notice, actual or constructive, of a prior inconsistent interest. Thus in our example, if C pays consideration and has no notice of the prior A-to-B transaction, then C will win. If C had actual notice of the deed to B, even if it is unrecorded, C cannot benefit from the statute. More importantly, a properly recorded deed furnishes constructive notice of B’s interest, thereby preventing C from being a GFPV. (Constructive notice can be record notice as here, or inquiry notice, i.e., knowledge of a fact that would put a reasonable person upon an inquiry that would result in actual notice. For example, seeing a ditch with running water on the land would put someone on inquiry notice to find out about an easement for the ditch.) As in the case of the race statute, B can protect B’s interests by recording as quickly as possible, which will prevent C from winning under the notice statute, because the recording provides constructive notice.
Some states, especially those of the mid-Atlantic and the old Northwest, have adopted race-notice statutes, under which a GFPV must win both the race to record and be without notice of the earlier transaction. Thus in our example, if A sells to B, and then A sells to C, who has no notice, and after that B records before C, nemo dat would still apply in favor of B. Why? Because although C would win under the “notice” prong of the test, C fails to win under the “race” prong because B recorded first. Of course as always, C can sue A, if that is feasible. Again, a transferee’s best bet,
170the oxford introductions to u.s. law: property
besides doing a title search, is to record as soon as possible. That way, the deed will be in subsequent purchasers’ chain of title and will furnish record notice to them.
The purpose of the recording acts in limiting the rights of owners vis-à-vis GFPVs is ultimately to make property more alienable. Prospective purchasers, who will be the new true owners, will be discouraged if they cannot be assured at reasonable cost that their grantor’s title is good. In general it is advantageous to owners that prospective purchasers have such assurance.
The recording acts also make a major exception to promote the alienability of property by someone who has won under the act. Consider the situation under a notice statute, in which A sells to B, and then B sells to C, who was without notice, and then D purchases from C knowing of the A-to-B transaction. Should D beat B? Taken literally, because D had notice of the A-to-B transaction, D is not a GFPV and so would not prevail under the act against B. But D’s grantor, C, who did not have notice, would beat B. The idea of the so-called shelter rule is that D should be able to shelter under C’s rights: Once C wins under the recording act, C can then transfer these rights to anyone without the transferee losing the protection of the act. So D wins against B by virtue of C’s rights. The shelter rule protects D by letting D step into C’s shoes. Otherwise C would have a hard time finding a buyer. (There is one exception to the shelter rule: If C transfers to A, the original grantor, then A cannot take advantage of the shelter rule, because a C-to-A transfer carries with it a danger of collusion. And discouraging someone like C from selling to someone like A does not diminish C’s potential market much).
In thinking about the chain of title, it is worthwhile to take the perspective of one searching the title records. Although it is true that adverse possession and some liens may not appear on the land records, the chain of title is meant to limit the search that grantees are responsible for. Consistent with this purpose, a deed needs to be not only recorded but also linked up with the rest of the chain of title. Consider the situation in which A sells to B who does not record, B then sells to C who records, and then A sells to D.
land transactions and title records 171
What happens under any of the types of statutes (assuming there is no tract index)? If D uses the grantee index to work backwards, D will find A, but working forward in the grantor index, D has no way to find C, because the intermediate link through B is absent. As far as the records are concerned, B is a “stranger to the title.” Again, D is not required to search the entire land records, which would be prohibitively expensive. Instead C is not counted as having recorded until Cs deed is linked up to A’s through a recorded chain. Otherwise Cs deed is termed a wild deed, and does not count as recorded vis- à-vis D. (Relativity applies, though, meaning that Cs deed is recorded for purposes of anyone claiming under C such as Cs successors in interest.) Again, the goal of the recordation system is to furnish cost-effective notice, even though, as we will see, problems do remain.
For personal property the exception to nemo dat in favor of GFPVs generally comes in through the Uniform Commercial Code (UCC) as enacted in all the states except Louisiana (which is a civil law jurisdiction). For example, if an owner entrusts a chattel to someone who deals in goods of that kind—for example, leaving a watch at a watch shop in a bailment for repair—a person purchasing the watch from the shop in the ordinary course of business will obtain good title even as against the original bailor-owner.3 This is probably a situation where the owner clearly can do more to avoid the problem than the subsequent good faith purchaser—the owner picked the bailee. But in many scenarios, deciding between innocent owners and GFPVs is very difficult in terms of comparative fault. By and large, U.S. law (of the various states) favors original owners more than does civil law or even English law, which tend to be relatively more solicitous of good faith purchasers. In U.S. jurisdictions, someone, even a GFPV, purchasing from a thief will lose to the original owner, unless enough
3. UCC § 2-403(2)–(3).
172the oxford introductions to u.s. law: property
time has passed, and the other requirements for adverse possession have been met.
Otherwise the UCC provides for limited circumstances in which a good faith purchaser for value can acquire title when the transferor has what is called “voidable title.”4 Suppose S sells a watch to P, who in turn sells to G. If P obtained the watch from S by fraud or using a check that bounces, P has a title that is voidable, because S can bring an action to have the transaction undone. (By contrast, “void” title is invalid for all purposes, without any need for one side to act to set it aside.) Under these circumstances, if P sells to G before S finds out about the defect in the S-to-P transaction, and G purchases from P in good faith (without notice of the problem) and for consideration, G obtains good title even against S.5 Of course, S can go after P, if possible. G must qualify as a GFPV in order to win; otherwise S beats G. Thus, if G had notice of the bad check, or if G did not pay consideration, G would not keep the watch. Sometimes a very low price in the subsequent (here, P-to-G) transaction can be evidence of bad faith on G’s part. In these voidable title situations of fraud and the like, it is harder to say in general who is more at fault: G and S are both innocent parties, but P (who is truly at fault) is typically not available, so one of them must bear the loss. Conceivably the loss could be split between the two, but this option is almost never used. Rather than try to allocate the loss based on fault or even a fifty-fifty split, the propertylike all-or-nothing approach is used instead. Here clarity and
predictability are usually at a premium.
Whether transactors qualify as GFPVs is largely governed by statutes such as the recording acts or the UCC. Occasionally, equitable principles fill in interstices left by these fairly
4.UCC § 2-403(1).
5.See Kotis v. Nowlin Jewelry, Inc., 844 S.W.2d 920 (Tex. Ct. App. 1992) (illustrating the operation of the UCC and upholding determination that G, to use our symbols, did not purchase from P in good faith).
land transactions and title records 173
comprehensive systems. Courts sometimes hold the original owner responsible for negligence for succumbing to a fraud, as where a fraudster tricks the owner into signing documents that turn out to be a deed.6 Some courts will allow a subsequent GFPV to prove that the original owner was negligent so as to estop the owner from denying the GFPV’s title. If the owner’s negligence led to justified reliance on the part of the GFPV, they hold that it would be unfair for the original owner to prevail against the GFPV. In economic terms, the original owner in such situations was the cheapest cost avoider—as determined on a relatively individualized basis. The application of estoppel here operates as a more fine-grained balancing of relative fault between the original owner and the GFPV and allows the GFPV to win even though the deed was a void one. This ex post approach is reminiscent of the equitable exceptions that courts first made to the original race-type recording acts. What the division of labor should be between hard-and-fast rules with ex ante clarity and ex post standards couched in terms of individualized justice runs throughout property law and beyond.
Not everything relevant to title is reflected in the land records. Some types of liens, such as “mechanic’s liens” for tradespeople who have worked on a building, and tax liens, might qualify the title without having to be recorded. And problems such as the incapacity of the grantor might invalidate the title, without any sign on the face of the records.
Adverse possession, which we encountered in Chapter 2, can also affect title without appearing in the record chain of title. If A acquires title by adverse possession against O and does not record, and then O sells the land to P who records, who owns the land? Even if P does a thorough search of the land records, A’s interest will not appear. Although it might seem that the policy of the recording acts should require that P win in such a case, nevertheless courts hold that A wins by adverse possession. P is required to protect P’s
6. See, e.g., Hauck v. Crawford, 62 N.W.2d 92 (S.D. 1953).
174the oxford introductions to u.s. law: property
interest against adverse possession by inspecting the land itself, in a form of inquiry notice. If someone with no apparent connection to the prospective grantor is occupying the land, the would-be purchaser must inquire further to find out whether the occupier has a claim based on adverse possession. If the statute of limitations has run, there can be no valid sale by the record owner. If the statute has not yet run, the record owner must remove the adverse possessor to provide marketable title: Remember that no one is expected to purchase a lawsuit. (Similar considerations of inquiry notice apply to easements by prescription, which we will take up in Chapter 8.)
Adverse possession also occurs in boundary disputes, which can cause special headaches for purchasers. Suppose A fences in a strip of B’s land and meets all the requirements for adverse possession (open and notorious, adverse, etc.) for the requisite period. Years go by, and the fence disintegrates. B then sells the land to C. As in our previous example, C is expected to inspect the land, but the fence is not there, so C has no way of knowing about the loss of the land to A. Nevertheless, adverse possession is robust enough that A still wins in such a situation, even though the idea of inquiry notice here is far-fetched.7 The law shows great solicitude that title by adverse possession is as valid as any other title, and successful adverse possessors like other owners are not required to mark their boundaries. One possible solution would be to require adverse possessors to record their interests. But notice that the only adverse possessors likely to do so would be those in bad faith (and those who have recently discovered the adverse possession). Depending on the purpose of adverse possession, this could be quite perverse.
Another method of preventing ancient history from being relevant in title disputes and searches is to cause them to be beyond the chain of title. Marketable title acts make unenforceable those interests that have not been recorded (or re-recorded) within a given time period working back from the present, typically the last forty years. This has the effect of eliminating constructive notice of
7. Mugaas v. Smith, 206 P.2d 332 (Wash. 1949).
land transactions and title records 175
any inconsistent unrecorded claims more than forty years old, so the old claims are eliminated by a transfer to a GFPV. And as noted earlier, title insurance is used to cover certain risks of defects in the title. These usually do not include adverse possession, which despite the fence example above, is usually best discovered by the purchaser inspecting the land.
Whether the problems of notice and the state of title are solved with bright line rules or vague standards or some combination of the two, it is important to remember that they raise the issue of relativity of title we saw in Chapter 2. Property rights are distinctively in rem, but questions of ownership are often presented in litigation between two identified parties. Such litigation can have what amounts to an in rem effect, however. If a GFPV or an adverse possessor (AP) “beats” the record owner, the GFPV or AP has title good against the world. But note that this does not mean that the original owner has nothing: The owner’s title still trumps everyone else’s. By the same token, relativity of title tells us that in the situation in which owner A is defrauded of real or personal property by B, who “sells” to C, and estoppel or the UCC tells us that C beats A, A still has better title than B, which leaves A with a claim for damages against B. Even when quiet title actions are brought to remove known clouds on title or resolve known problems (e.g., adverse possession), these actions, although in theory in rem, tend to involve identified competing claimants. Only in registration systems does a registrar give conclusive title against all claims both known and unknown. For this reason the registrar can be thought of as standing in for the (in rem) public and correspondingly, registration countries (notably but not exclusively those following the German system) tend to have a strict numerus clausus (Chapter 5). Registrars can pronounce on title more easily if they need not incur the information costs of figuring out how idiosyncratic interests fit into the array of interests relating to the land in question.8
8.See Benito Arruñada, Property Enforcement as Organized Consent, 19 J.L. Econ. & Org. 401 (2003).