Interpreting the census populations and their potential impact on prices
In this topic, I will attempt to analyze the census counts. I will include examples from the USA, from South Africa and Southern Rhodesia for illustrative purposes only due to the discussion in the Baldwin sale and because I consider it relatively representative of those markets where TPG is not preferred, which by my reckoning only includes the United States, South Africa and maybe Canada. The format I will use is to ask and answer what I believe to be the relevant questions. Some of this will be a “rehash” of what has been discussed here before (primarily by me).
In this commentary, I will cover a few specific coins but if anyone else has an interest in any others, I will give it a go to provide my interpretation.
Question #1: Why are the census counts generally so low for most coins?
Answer: Absent evidence to the contrary, the primary reason is due to 1) collectors outside of the markets listed in my introduction preferring ungraded coins and 2) The coin’s market or at least catalog value. The second one is invariably true even in the United States except where an alternate explanation exists. An example is the NGC and PCGS Registry Sets. This is why the census count for a coin like the USA 1983-P Washington Quarter in MS-65 is so high; in the hundreds. Otherwise, there is no reason to pay a grading fee for a coin which is “only” worth $30.
So what I am telling and have repeatedly told everyone here is that simply because a coin has a low census count does not make it “rare” or even “scarce”. Without the proper context, the data doesn’t mean anything. South African Union coins are scarcer than most and are as scarce now as they were when the census counts were much lower five and ten years ago. It is just that we did not collectively know where most of these coins were or if they actually existed. But it is really only for the two reasons I give that they have increased as much as they have and like I have said many times, invariably, there are still going to be more or a lot more of these coins available to be graded across the grade spectrum. This should also be the normal expectation elsewhere unless a specifc reason exists to invalidate it.
Last edited by jwither; 02-10-13 at 21:49.
Question #2: To what extent are the census counts distorted by duplicates, including cross overs between NGC and PCGS?
Answer: First, it is going to primarily depend upon the absolute value of the coin and the price spread between two different (generally MS) grades. In the example I gave above of the 1983-P Washington Quarter, the last sale of this “key” date on Heritage was for slightly over $100. The one prior to that, $293. Between these two sales, the PCGS census count for MS-66 soared from 114 to 217. I know this because I tracked it. In MS-65, it is now 498. Because the price spread was a lot wider before, maybe there are a large number of duplicates in MS-65 but I doubt it. There are probably a lot more in MS-66 since an MS-67 is the “conditional rarity” (with 6) and the price is likely in the mid to high four figures (as a guess). But even so, there is no reason to believe that most MS-66 or even a disproportionate percentage of them are duplicates.
In a recent PCGS forum post, contributor “TDN”, the co-owner of the 1794 SP-66 dollar which recently sold for $10MM, mentioned that eight MS are known. The combined census lists 14 at this time, including the SP-66 and two other MS-66. Apparently, there are two additional “gems” that actually exist (presumably the MS-66) but we can know from his comments that six duplicates are in the census today. This makes sense since any MS example is worth over $1MM today.
The second reason is the likelihood of the coin upgrading. Simply because the price is “high” or because the price spread is “wide” between two grades does not mean that the census counts include any duplicates, much less a large number. If the probability is low or not that high, then there is no point to “crack out’ the coin and resubmit it unless the incremental value is disproportionate to the cost of multiple submissions. Only an ignorant or uninformed collector or “investor” is going to randomly or repeatedly do so. It happens but doing so is far more likely to be a waste of money. Whether the probability is “low’ or “high” depends upon the grading standards in use at the time and where any particular coin falls within its grade because as I have repeatedly stated, all coins with the same grade are not equal.
This is what some or many apparently seem to believe with the Union census. I believe the number of duplicates is either minimal or nonexistant and where it does exist, it is frequently crossing a PCGS to an NGC holder. And even with the coins where it makes some sense, I still do not believe absent specific evidence that the number is even disproportionately significant. In all likelihood, because of the price difference, the most likely candidates for duplicates within Union and ZAR are AU-58 and not any MS grade. For Southern Rhodesia, we cannot know if any exist but I do not see any reason to believe so absent specific evidence to the contrary either.
Because of their value and the size of the current counts, the only market where I believe an outsized number of duplicates exist is with United States coinage. When I recently checked the combined census, the 1795 half dollar listed 70 in MS. I suspect that it might proportionately have more duplicates than the 1794 dollar even though it is worth much less. I suspect it first because I do not believe this many of a coin like it would have survived in this condition and second because it is less likely that resubmissions would be recognized as the same coin, as I think would happen with the 1794 dollar. When an upgrade will result in a “large” price increase, purportedly NGC and PCGS are more conservative in giving it. This makes sense because they have some exposure from the grading guarantee.
With some coins, I have heard that the actual number is a fraction of the count but these coins are all worth a lot of money, much more than practically all Union or ZAR. I have heard of some US coins being submitted up to 10 times. This does happen but for anyone who thinks this is more common than I do, they also might want to consider that whoever did it potentially ended up with a lower grade versus a higher one.
Last edited by jwither; 03-10-13 at 18:50.
Question #3: What is the significance of a “conditional rarity”?
Answer: It depends upon the census distribution, the price of the coin and the popularity of the coin or series.
The first conclusion that I derive from the data is that “conditional rarities” below MS-66 or MS-67 are usually not considered that significant with US collectors because most are at or above this cut-off. And even where they are not, I have never heard of anyone making a big deal out of those which are in the same grades as most Union or ZAR, not even once.
The second conclusion is that, the greater the count right below the “conditional rarity” and the larger the total census count, the more significant. This is why I once stated that the Union 1923 NGC MS-67 BN 1D should have the greatest premium in this series. The last time I checked, there were 41 in MS-66 so I do not think the price should be high (probably $200-$300 at most versus $500-$600) but the premium should be higher versus others. The lopsided proportion of the rest, a minimal or nominal premium and nothing more. Logically speaking, what is such a big deal about owning the “top pop” when there are also only a small number right below it (or maybe even just one) and why would anyone pay an outsized premium for such a coin? I mean, who cares? This especially applies to coins that are extremely scarce such as the 1925 florin. In the United States where TPG originated, this distribution is particularly important because of availability but also because of Registry Set competition.
The third conclusion is that lower priced coins generally have greater premiums than more expensive ones. This presumably exists because if it did not, the coins would either be completely unaffordable or buyers will find better value elsewhere.
Taking the combination of these factors, this is why US circulating moderns have the greatest premiums while those for “key” dates in the “perennial collector favorites” are relatively low..
With this context, now it should be more apparent why the premiums of especially most better Union coins are absurd; the Registry Set is not a factor in collecting at all from what I can see. No, this isn't the only reason why all or even most high grade US coins sell for substantial premiums or a lot of money but it is a big one for many in the same price range as most Union, ZAR or non-US coins generally. Though NGC has many sets for Union and ZAR, the number of participants is immaterial, like a handful or even less for each one. In the US, the PCGS Registry is a lot more popular than NGC and especially so for circulating moderns such as the 1983-P Washington Quarter described above. While I have no interest in such competitions, apparently many others like them or else some series would not have dozens or even hundreds of participants. In this (illogical) context, it makes "sense" given the number of collectors in total and registry participants that some will pay absurd prices for the "best" specimens, even for coins that are otherwise as common as a grain of sand on the beach.
Last edited by jwither; 04-10-13 at 17:07.
Question #4: Why are the census counts so much higher with some coins than others, even in a market like Southern Rhodesia?
Answer:To answer this question, I believe you need to understand collector preferences.
The combined census for the 1953 Rhodes centenary crown is 88 (85 in MS) circulation strikes and 31 more in PR. This coin is the most popular because of its size and as the only crown. It is as simple as that. It is presumably also the most common but the census counts are primarily a function of its popularity and not its relative scarcity. In other words, if the coin was as popular as the rest of the series, the census count would be much lower even though its scarcity would not change at all. I happen to own an NGC MS-63 even though I do not collect this series otherwise.
What I just described, the same is true of the Bolivian Republic decimals I collect where the crown sized 1872 Boliviano has a census count in MS of about 30. This denomination is the most (and only relatively) popular one from this series because collectors generically like big “old” silver coins, especially if “cheap” and in “high” grade. Since I do not believe there are anywhere near 88 or 30 “serious” collectors of these two series, most of these coins are almost certainly owned by those whose focus is elsewhere.
The 1932 proofs have census counts from the high 20’s to the high 30’s and I have seen many ungraded sets sold at auction with most more likely remaining ungraded today than not. I suspect that some or many of these sets are also bought by those who do not primarily collect this series, probably mostly those who collect British Commonwealth generically. For those who do not collect either, the actual reason I do not know but I suspect it is partly because some think the mintage of 492 is low (it isn’t) and it is the only KGV proof set from this market or at least the only one available. It is possible that the attrition on this set is higher than I believe but even if only 50 remain (the odds which I would rate as effectively nil), this still isn’t that low for an obscure market like this one and yes, it is obscure because while “some interest” does not mean “none”, it is still lower or much lower than others with a collector base of any consequence.
The 1946 2/ which was mentioned in “Scarce Coin Watch”, this coin has a census count of 12 in MS (20 in total). Whether it is scarcer or more common than most others in the series, we do not know or at least, I do not. The likely reason why this coin has this count and more than most others is that it is a “key” date but one which is probably not as scarce as is generally believed by those who have bought it, just like some of the low mintage Union KGVI 2/ and 2/6. The census counts for the 1939 and 1954 2/ (also “key” dates) are much lower. Even though I believe the latter two have a higher supply than indicated now (much larger proportionately), if my explanation is at least directionally correct, then I would expect to see more of these others (especially in better grades) if the actual relative scarcity is in line with current prices. The same applies to the 1954 1D Geejay mentioned which has an MS count of six. This coin is presumably at least somewhat scarce based upon the price but I see no basis from the census data alone to believe it is really any scarcer than most Union KGV bronze such as the 1927, 1928 and 1933 1D.
With the rest of the dates, the only firm conclusion I believe that can be drawn from them is that those which are “outliers” are not scarce at all. An example would be the 1937 shilling with a census count of 24, 16 in MS and several “gems” including an MS-67. I suspect that many exist in high or MS grades, maybe in the hundreds.
For those with the lowest census counts today, I can also come to the logical conclusion that though we cannot know with certainty that the actual availability is higher or much higher than it currently shows, it is highly likely. I can deduce this by making a common sense evaluation of the census counts from elsewhere. To use but two, the 1746, 1751 and 1752 Mexico ½ real show 17, 24 and 13 in MS. A disproportionate number are graded MS-63 but this still doesn’t mean there are many if any duplicates. There isn’t much to be gained with multiple submssions because the price difference is nominal. And given what I have said elsewhere on this subject, I still believe all of these coins have more available.
For the “key” date of the entire Southern Rhodesia series, the 1939 2/, it has zero MS listed today while the 1771 Peru 2R which I own in MS-62 has four. Except for the 1758, 1761 and 1771, in 10+ years, I have never seen any other Peru 2R MS (graded or not) for sale anywhere and only a handful more in AU. Some of the other dates included as plate coins in my Gilboy reference manual are “high” grade or MS, but many others throughout the series which he lists as supposedly common, the quality of the coins he was able to use as examples are terrible. This indicates to me that these are actually scarcer or much scarcer than he claims, or else he should have been able to find better examples.
So the question is this one. Given that one series is from the 20th century and the other from the 18th, the difference in economic development, in communication and in travel, are we to believe and does it remotely make any sense at all that more than a few Southern Rhodesia (or Union) circulation strikes are actually scarcer in better grades or in total than these coins? If so, based exactly upon what? Even if Southern Rhodesia were disproportionately melted that still does not remotely get us from the original mintage figures to the current census data, just as it did and does not for South Africa Union.
Last edited by jwither; 09-10-13 at 20:48.
Question #5: Where does most of the demand for Southern Rhodesia TPG coins likely originate?
Answer: I believe most of it comes first from South Africa and second, from the United States. Given the low census counts in most instances, it might almost always be either or both. South Africa has geographic proximity to Zimbabwe and is a logical choice for collectors in your country as a second series. I suspect that just like US collectors of Canadian coins, there are some in your country who have this preference for the same reason. A second reason may be that some in South Africa possibly believe they can replicate the economic windfalls which were realized earlier with ZAR and Union.
The United States as the second alternative because I believe this country disproportionately is the source of demand for most graded non-US coins, regardless of what country we are talking about. I do not believe that generic British Commonwealth collectors have any evident preference for TPG and the same applies to the few (likely European) collectors that remain in Zimbabwe now, the latter which I believe is almost none.
Question #6: For a market like Southern Rhodesia, how do current prices reflect the census counts and grade distribution?
Answer: I do not believe they do much if at all and not just for Southern Rhodesia but practically every coin series where TPG is not preferred. In the recent Baldwin’s sale, I do not see that most actual or prospective buyers would have established their bids using this data. It doesn’t make any sense because most likely do not care about graded coins, the counts are so low and because a coin in a particular grade is sold so infrequently. If they do, then the collector base must be even less than I believe it to be.
Last edited by jwither; 15-10-13 at 15:25.
Question #7: Given the results of the Baldwin’s sale, what do the current and future census counts likely mean for Southern Rhodesia coin prices going forward?
Answer: Absent an explosion in demand for “rare” coins generically or at least British Commonwealth, I do not see that it will have much of an impact at all (in either higher or lower prices) unless the counts turn out to be higher as I expect. If the actual supply turns out to be as large as I believe it will be in many if not most instances, it goes without saying that it will negatively impact pricing. These comments, I also apply to practically everywhere else, including the other series I collect.
If by some accident of history, the current counts are mostly complete or near it, then the obstacles I described in my recent commentary titled “A comparison between the market characteritics of the United States and South Africa” should apply. As with the South Africa and Mexico examples I used, I do not see the few collectors or “investors” of these coins paying “moon money” when they are not available to be bought, regardless of how rare they turn out to be. To take the 1939 2/ as the “key date” of the entire series, if there really are only a handful and none in MS as now and “everyone knows it”, I still see no basis to believe that this coin is going to be worth (for example only) 10 or 20 times its current value (in current prices), much less more. There simply aren’t enough collectors who are likely to care about this coin or this series enough who “must” have it for that outcome to last except maybe temporarily. Realistically, this allows for some price appreciation but likely not out of the ordinary.
On the size of the collector base, I cannot tell you exactly what it is but I suspect it is some undeterminable fraction of those who collect Union. In one email exchange, a collector in your country (who many of you know) once told me that maybe 120 “serious” collectors exist for Union. (He provided no estimate for ZAR.) I have no clue where that number came from but it sounds reasonable to me if by “serious”, we are talking about someone who can and is willing to spend as much as $500 or R5,000 per month on this series. Given price trends since year-end 2011, it would not surprise me if the current actual number is less because 120 buyers who are actually spending this much would appear to be more than enough to support higher or much higher prices.
Using this 120 as a baseline, I would then guess that maybe Southern Rhodesia has in the vicinity of a few dozen or maybe at most slightly more with the qualification that I doubt they are spending the equivalent of $500 per month because the coins do not appear to be available at this time. I did not see an auction report for the Baldwin’s sale but I suspect that the majority of “serious” collectors did bid given that few opportunities exist to buy these coins at one time in any quantity.
The other obstacle that I see is that they are really not that cheap given their potential or actual availability and the current or likely demand. This is subjective of course but I think I can provide much better evidence to support my claim than those who disagree with it.
First, let’s compare the prices of the 1939 Baldwin’s proof set which sold for $11000+ and the 1937 Heritage set which sold a year ago for $4700. The second set cost about as much as the concurrent Baldwin sale of the Union 1938 at 2800 GBP but with six coins instead of eight. Heritage claims that only 20-25 1937 sets were actually sold versus the mintage of 40. I am dubious of this lower estimate because of the sale frequency but even so, I do not see that this set is more desirable than the Union 1938. I consider it inferior unless there are far fewer than even 20-25 because it really is not that hard to buy at all.
The relative desirability of the 1939 set will depend upon how many of the 10 remain and the Baldwin’s price implies it is currently worth somewhat less than the equivalent denominations in the 1934 Union set. I suspect most of these 10 sets are available somewhere and using this assumption, I would rate it somewhat superior to the Union 1934 but mainly because it is harder to actually buy and it may be the most prominent proof set of the entire series. But in any event, I do not see that either of these two sets represent compelling value versus these alternatives (or others like them) and I do not see that their future financial prospects are any better either.
Next, let’s look at the prices of the 1939 2/, 1946 2/, 1954 2/ and 1954 1d from the recent Baldwin’s sale and compare them to equivalent Union. The 1939 2/ in GEF (AU-55 to MS-63 in my experience) sold for about $850. If it actually is an AU-55 or AU-58, there aren’t many Union at all that sell for this much and only a handful from KGVI. I suspect that it isn’t actually any scarcer than the Union KGVI (such as the 1938 or 1946 2/) and if it isn’t, I do not see any compelling relative value there either in AU. It only is if this coin is actually an MS that I see it. The only alternative argument that I can make in favor of it is that it is the “key” circulation strike for the whole series while the other two compete with many others.
The 1954 1D which sold for about $750, I do not see that the current count of six makes it actually any scarcer and certainly not more desirable than KGV bronze such as the 1927, 1928 and 1933 pennies. As a “gem” I presume that its potential TPG grade is an MS-65 or maybe even better but with this grade, it might sell for a comporable price to these Union counterparts even now. If so, where exactly is the room for substantial future appreciation?
The 1946 2/ I already discussed under “Scarce Coin Watch”. With a current MS count of 12 and a likely actual one which is some multiple of it, I do not believe that its actual scarcity justifies the most recent price of $1200, even in a grade like MS-65. Given this likely availability and the difference in popularity between the two series, I do not see that it should sell for more or even the same price as any of the Union “key” 2/ dates in equivalent grade either.
For the 1954 2/, I think the most recent price of $1300+ is a lot more reasonable than the 1946 2/ because I have seen it much less often. I do not believe it is overpriced but I’m not convinced it is a bargain or a compelling “investment” at that price either.
Last edited by jwither; 15-10-13 at 15:29.
Question #8: Taking into consideration the evidence and assumptions included above, what is the most reasonable conclusion to draw on Southern Rhodesia as a series from what we see in the census today?
Answer: The short answer is to collect it if you like it but don’t count on any economic windfall by doing so. I see no compelling argument for such a belief whatsoever.
The slightly longer response is that first, these coins are likely more common or a lot more common than is apparent from the census today, even if they are still “scarce” or “rare” versus others.
Second, there is likely a minimal or extremely limited collector base for it now and I see no compelling argument that it will increase to any substantial extent in the future. If anything over the longer term, I expect it to shrink because Rhodesia and Southern Rhodesia no longer exist; numismatically speaking Zimbabwe I do not see as equivalent because I do not believe collectors of either disproportionately collect the other at all. As few collectors of Zimbabwe coinage likely exist today, I do not see that the likely predominantly African collector base does or ever will show any substantial interest in Southern Rhodesia, just as this does not exist in South Africa today.
Third, we do not know how much more existing or future collectors can or will pay, but there isn’t a compelling argument that I can see that they will do so substantially absent the same outcome in coins generally or least in similar ones specifically. Some yes or maybe but not to the extent as with Chinese, some other Asian contries or Latin American coinage.
What this means from a practical standpoint is that there are almost certainly much better options for the “investor” elsewhere. I am not interested in Chinese coins and know nothing about the market characteristics of other Asian countries except a little for Japan and Korea. I am not sure what the relative value proposition is today between these alternatives I list and Southern Rhodeia. What I do know is that coins with little or no demand are "cheap" for this reason while coins which are far more common such as the United States are a lot more expensive.
Last edited by jwither; 15-10-13 at 17:17.