# Rate of return to investment in American antique furniture.

I. Introduction

Most research on the market for "collectibles" has focused on their potential for long term price appreciation. Formal and informal estimates of the rate of return on visual art, especially paintings, are common |1; 2; 4; 6; 10~. An interesting exchange on bottled wine as an investment is provided by Krasker |8~ and Jaeger |5~. Recently, Ross and Zondervan |9~ analyzed the investment market for Stradivarius violins.

The research strongly suggests that paintings as a group are a poor long-term investment. Ross and Zondervan |9, 539~ conclude the Stradivarii ". . . may give a return in excess of the long-term real rate of interest of 2.5 per cent if taxes are absent and the user benefits are significant," but they are unable to establish a clear indication of user benefits. A possible exception to these negative assessments is bottled wine. Jaeger |5, 590~ shows that correcting Krasker's |8~ storage cost estimate and taking a time period that bridges recessions (to which all wine prices are very sensitive) yields a substantial premium rate of return on fine wines over the T-bill rate. Her explanation |5, 584~ is that bottled wines improve in quality up to an (uncertain) peak year, which can range from ". . . 20-40 years in the case of some reds, while a few refuse to retire at 65," after which their market value drops sharply; we see here an apparent influence on the wine market of user benefit.(1)

This paper will extend the above research by estimating rates of return on American antique furniture (hereafter, "antiques").

II. Antique Furniture Prices

Data

By far, the most extensive data base of antique prices is provided by Ralph and Terry Kovel |7~. Their mail surveys and personal research yield almost a thousand antique furniture price quotes yearly. In the Kovel's data base, "Prices are the actual asking price, although the buyer may have negotiated to a lower figure . . . each price is one you could have paid for the object" |7 (1987), p. v.~. In other words, these are raw data, directly from the showroom. But the Kovels also omit items of no interest to collectors, such as "used" furniture and old furniture in bad condition. They omit survey responses with ambiguous, incomplete or inconsistent data. And they omit outliers--extremely high prices due, in their opinion, to "'auction fever'" |7 (1990), iii-v~. For the Kovels, the behavior of typical American collectors defines "antique" furniture.

Regarding opportunity costs, we follow precedent. We compare the investment "yield" implicit in the behavior of antique prices with the yield on U.S. Treasury 90-day bills |3~.

Data Management

Generally, unlike Stradivarius violins, bottled wines, etc., one cannot trace the market value of a given piece of antique furniture over time. The exceptions are those very rare, expensive items whose auction prices are unrepresentative of the rest of the market. Moreover, even if one could tag a piece of furniture and follow its market history, the result (as we infer below) would be an unreliable estimate of the huge market for antiques in general. Antiques' heterogeneity, as to place of sale (they are expensive to ship) and individual physical characteristics (requiring personal inspection) causes substantial and unpredictable variations in price. An antiques 'investor' who concentrates on a narrow genre and type is accepting above average risks; such dealing is really speculation. The wide diversity of items carried by the typical retail dealer is partly an effort to minimize these risks.

Fortunately, similar or identical furniture appear at estate auctions, in dealers' showrooms and in advertisements year after year. We may gain insight into antiques' investment potential by constructing 'portfolios' of various general types of antiques. These portfolios can be assembled from the Kovels' |7~ antiques data base. It is only necessary to construct a price index for certain broad categories.

In this study, we construct indices of the price history of the living room chair; the dining room chair, table, buffet, cabinet; bedroom dresser (or vanity), chest of drawers, and bedstead. These eight general furniture types are each filled from the very similar sub-types identified among the Kovels' |7~ price data. The actual procedure mimics that for any price index. First, we determine the mean price, in a given year, of a particular kind of furniture, e.g., "candlestands" as a sub-type within the general type of living room tables:

|Mathematical Expression Omitted~

This is the mean price, |Mathematical Expression Omitted~, of the ith sub-type (e.g., candlestands), within the jth general type (e.g., Living Room tables, here) for which we find in Kovel |7~ K price quotes(2) in the year t. These mean prices are then fixed-weighted using the aggregate of the K values of the given (or ith) sub-type as reported by the Kovels |7~ over the full 20 year period:

|Mathematical Expression Omitted~

Using the resulting values of |Mathematical Expression Omitted~ and N, it is a simple matter to combine the n sub-type values (i = 1, . . ., n) common to a given jth general type, into a fixed-weighted index (or average) of the general type,

|Mathematical Expression Omitted~

from which we have the index price, |P.sub.jt~, of general type j furniture (j = 1, . . .,8) in period (t = 1, . . ., 20).

TABULAR DATA OMITTED

After aggregating and fixed-weighting, we have a portfolio based on a specific product mix defined for the entire 20 years by the fixed-weights.(3) From these weighted means, each year we have 8 antique index prices, with a total of 8 x 20 = 160 such prices over the 20 years. These prices are the expected prices available to antiques investors, although much of the risk-generating price variance has been suppressed.

III. Rates of Return

Following |9~, we estimate the capital gains rate of return on a general type of antique furniture from

|P.sub.t~ = |P.sub.0~|e.sup.rt~.

We define |P.sub.t~ and |P.sub.0~ as period 'endpoint' prices, using the mean prices of the first three and last three years of the 20 year span, yielding a time-base of 18 years. The mean price of the four middle years is used as an end 'year' when calculating beginning and ending 9 year estimations. Using these end points, r is a continuous compound-interest statistic.

The results of these calculations are shown in Table I. We first observe that the yields vary across types. For example, for the full 20 years (row 1) the highest rate of return (row 1, column 7) was 11.9 per cent while the least return (row 1, column 3) was 3.5 per cent. Despite our fixed-weight aggregation of sub-types into more diversified portfolios, the remaining influence of antiques price variance is substantial.

The price variance would be of little interest if rates of return exceeded opportunity cost. The estimated annual rate of return on all antique types, collectively, for 1967-1986 is shown in Table 1 (row 1, column 1) as 6.97 per cent. The corresponding 90-day T-bill yield |3~ (row 1, column 10) is 7.31 per cent. Even if retail and wholesale antique prices were alike, the net return from buying and holding a broadly inclusive portfolio of antiques for the 20 years would have been minus 0.34 per cent.

The time period makes a difference. In the decade from 1967 through 1976, the inclusive antique portfolio earned 8.24 per cent while T-bills' yield was only 5.66 per cent (row 2, columns 1 and 10), for a 2.58 per cent premium. As with many collectibles, however, the wholesale discount is significant; for antiques, from 20 to 50 per cent |7 (1987), v~. The 2.58 per cent premium is offset or more than offset by the wholesale discount.

As noted above, the more specialized the portfolio, the greater the price variance over time and the greater the opportunity for gain or loss. By choosing the right antique type, instead of a more diverse portfolio, one could have done better. For example, a portfolio of beds bought and held during 1967-1977 yielded 15.3 per cent (row 2, column 7) when T-bills were yielding 5.66 per cent. The largest wholesale discount of 50 per cent would still have yielded almost 3 per cent net. But, as with all relatively specialized portfolios, ab initio a sub-set like this would have been riskier, especially relative to financial instruments.

During the second half of the two decade span, opportunities for gainful investment in antiques deteriorated substantially. The antiques' overall rate of return fell from 8.24 per cent in the first decade to only 5.71 per cent in the second decade (rows 2 and 3, column 1), while T-bill rates were soaring (rows 2 and 3, column 10) from 5.66 to 8.99 per cent. Every portfolio depicted in the table, even beds, would have yielded a clear loss.

IV. Conclusion

The results in Table I are generally consistent with those of previous research. Studies of the monetary rate of return on investments in collectibles, through long-term price appreciation, typically conclude they are a poor investment relative to financial instruments.

Since collectibles, in fact, are more than mere financial instruments to their owners (who otherwise presumably would be collecting financial instruments, instead), one suspects we are measuring collectibles' rates of return against the wrong criteria. Along with pecuniary gains, we ought to be including nonpecuniary user benefits. Only in one case, that of bottled wine, does the pecuniary yield appear to reflect the natural increase in user benefit at the final imbibing.

The problem is that user benefits, from daily association with an aesthetic object, are not readily disentangled from the object's financial yield. If use and ownership were separate, the user paying a rent to the owner for exclusive daily association with the collectible and the owner getting, in addition to the rent, the financial yield from any price appreciation, then the market would reflect the total return--rent plus capital gain--from the collectible. One suspects the mean value of this total return would easily stand comparison with yields on financial instruments.

Such rental arrangements seem very rare. At least, we have no public data base of the kind. There is the problem of agency. The principal hindrance to such a market may be the lack of motivation in the renter to preserve and protect the aesthetic object as would an owner-in-possession. Buying collectibles to rent out, therefore, may simply be unprofitable.

Paul Graeser Northern Illinois University DeKalb, Illinois

1. The benefit from consuming a great vintage wine is a user benefit. The wine's value as a collectible inheres in its age and vineyard, which are aesthetic qualities.

2. In six per cent of the sub-type cells, |K.sub.ijt~ = 0. The fortran program written for data management deals with these empty cells, thus avoiding the obvious bias. (I am indebted to Susan Porter-Hudak for this useful program.)

3. The method is similar to computing the well known Consumer Price Index. One difference is that the antiques index uses the entire 20 year period to determine the weights, instead of the less reliable Laspeyre's single base-year weights. Also, the antique "product" itself does not depreciate. Indeed, the antiques market is essentially a circulating stock: vendor offerings of antiques, unlike offerings of typical consumer goods, collectively change little or none over time.

References

1. Anderson, Robert C., "Paintings as an Investment." Economic Inquiry, March 1974, 13-26.

2. Baumol, William J., "Unnatural Value: or Art Investment as a Floating Crap Game." American Economic Review, May 1986, 10-14.

3. Federal Reserve Bulletin, 1970-1988. Washington, D.C.: Board of Governors of the Federal Reserve System.

4. Frey, Bruno S. and Werner W. Pommerehne, "Is Art Such a Good Investment?" The Public Interest, Spring 1988, 79-86.

5. Jaeger, Elizabeth, "To Save or Savor: The Rate of Return to Storing Wine." Journal of Political Economy, No. 3, 1981, 584-92.

6. Keen, Geraldine. Money and Art: A Study Based on the Times-Southeby Index. New York: G.P. Putnam's Sons, 1971.

7. Kovel, Ralph and Terry Kovel. Kovels' Antiques and Collectibles Price List. New York: Crown Publishers, annually 1968-1987.

8. Krasker, William S., "The Rate of Return to Storing Wines." Journal of Political Economy, No. 6, 1979, 1363-67.

9. Ross, Myron H. and Scott Zondervan, "Capital Gains and the Rate of Return on a Stradivarius." Economic Inquiry, July 1989, 529-40.

10. Stein, John P., "The Monetary Appreciation of Paintings." Journal of Political Economy, October 1977, 1021-35.

Most research on the market for "collectibles" has focused on their potential for long term price appreciation. Formal and informal estimates of the rate of return on visual art, especially paintings, are common |1; 2; 4; 6; 10~. An interesting exchange on bottled wine as an investment is provided by Krasker |8~ and Jaeger |5~. Recently, Ross and Zondervan |9~ analyzed the investment market for Stradivarius violins.

The research strongly suggests that paintings as a group are a poor long-term investment. Ross and Zondervan |9, 539~ conclude the Stradivarii ". . . may give a return in excess of the long-term real rate of interest of 2.5 per cent if taxes are absent and the user benefits are significant," but they are unable to establish a clear indication of user benefits. A possible exception to these negative assessments is bottled wine. Jaeger |5, 590~ shows that correcting Krasker's |8~ storage cost estimate and taking a time period that bridges recessions (to which all wine prices are very sensitive) yields a substantial premium rate of return on fine wines over the T-bill rate. Her explanation |5, 584~ is that bottled wines improve in quality up to an (uncertain) peak year, which can range from ". . . 20-40 years in the case of some reds, while a few refuse to retire at 65," after which their market value drops sharply; we see here an apparent influence on the wine market of user benefit.(1)

This paper will extend the above research by estimating rates of return on American antique furniture (hereafter, "antiques").

II. Antique Furniture Prices

Data

By far, the most extensive data base of antique prices is provided by Ralph and Terry Kovel |7~. Their mail surveys and personal research yield almost a thousand antique furniture price quotes yearly. In the Kovel's data base, "Prices are the actual asking price, although the buyer may have negotiated to a lower figure . . . each price is one you could have paid for the object" |7 (1987), p. v.~. In other words, these are raw data, directly from the showroom. But the Kovels also omit items of no interest to collectors, such as "used" furniture and old furniture in bad condition. They omit survey responses with ambiguous, incomplete or inconsistent data. And they omit outliers--extremely high prices due, in their opinion, to "'auction fever'" |7 (1990), iii-v~. For the Kovels, the behavior of typical American collectors defines "antique" furniture.

Regarding opportunity costs, we follow precedent. We compare the investment "yield" implicit in the behavior of antique prices with the yield on U.S. Treasury 90-day bills |3~.

Data Management

Generally, unlike Stradivarius violins, bottled wines, etc., one cannot trace the market value of a given piece of antique furniture over time. The exceptions are those very rare, expensive items whose auction prices are unrepresentative of the rest of the market. Moreover, even if one could tag a piece of furniture and follow its market history, the result (as we infer below) would be an unreliable estimate of the huge market for antiques in general. Antiques' heterogeneity, as to place of sale (they are expensive to ship) and individual physical characteristics (requiring personal inspection) causes substantial and unpredictable variations in price. An antiques 'investor' who concentrates on a narrow genre and type is accepting above average risks; such dealing is really speculation. The wide diversity of items carried by the typical retail dealer is partly an effort to minimize these risks.

Fortunately, similar or identical furniture appear at estate auctions, in dealers' showrooms and in advertisements year after year. We may gain insight into antiques' investment potential by constructing 'portfolios' of various general types of antiques. These portfolios can be assembled from the Kovels' |7~ antiques data base. It is only necessary to construct a price index for certain broad categories.

In this study, we construct indices of the price history of the living room chair; the dining room chair, table, buffet, cabinet; bedroom dresser (or vanity), chest of drawers, and bedstead. These eight general furniture types are each filled from the very similar sub-types identified among the Kovels' |7~ price data. The actual procedure mimics that for any price index. First, we determine the mean price, in a given year, of a particular kind of furniture, e.g., "candlestands" as a sub-type within the general type of living room tables:

|Mathematical Expression Omitted~

This is the mean price, |Mathematical Expression Omitted~, of the ith sub-type (e.g., candlestands), within the jth general type (e.g., Living Room tables, here) for which we find in Kovel |7~ K price quotes(2) in the year t. These mean prices are then fixed-weighted using the aggregate of the K values of the given (or ith) sub-type as reported by the Kovels |7~ over the full 20 year period:

|Mathematical Expression Omitted~

Using the resulting values of |Mathematical Expression Omitted~ and N, it is a simple matter to combine the n sub-type values (i = 1, . . ., n) common to a given jth general type, into a fixed-weighted index (or average) of the general type,

|Mathematical Expression Omitted~

from which we have the index price, |P.sub.jt~, of general type j furniture (j = 1, . . .,8) in period (t = 1, . . ., 20).

TABULAR DATA OMITTED

After aggregating and fixed-weighting, we have a portfolio based on a specific product mix defined for the entire 20 years by the fixed-weights.(3) From these weighted means, each year we have 8 antique index prices, with a total of 8 x 20 = 160 such prices over the 20 years. These prices are the expected prices available to antiques investors, although much of the risk-generating price variance has been suppressed.

III. Rates of Return

Following |9~, we estimate the capital gains rate of return on a general type of antique furniture from

|P.sub.t~ = |P.sub.0~|e.sup.rt~.

We define |P.sub.t~ and |P.sub.0~ as period 'endpoint' prices, using the mean prices of the first three and last three years of the 20 year span, yielding a time-base of 18 years. The mean price of the four middle years is used as an end 'year' when calculating beginning and ending 9 year estimations. Using these end points, r is a continuous compound-interest statistic.

The results of these calculations are shown in Table I. We first observe that the yields vary across types. For example, for the full 20 years (row 1) the highest rate of return (row 1, column 7) was 11.9 per cent while the least return (row 1, column 3) was 3.5 per cent. Despite our fixed-weight aggregation of sub-types into more diversified portfolios, the remaining influence of antiques price variance is substantial.

The price variance would be of little interest if rates of return exceeded opportunity cost. The estimated annual rate of return on all antique types, collectively, for 1967-1986 is shown in Table 1 (row 1, column 1) as 6.97 per cent. The corresponding 90-day T-bill yield |3~ (row 1, column 10) is 7.31 per cent. Even if retail and wholesale antique prices were alike, the net return from buying and holding a broadly inclusive portfolio of antiques for the 20 years would have been minus 0.34 per cent.

The time period makes a difference. In the decade from 1967 through 1976, the inclusive antique portfolio earned 8.24 per cent while T-bills' yield was only 5.66 per cent (row 2, columns 1 and 10), for a 2.58 per cent premium. As with many collectibles, however, the wholesale discount is significant; for antiques, from 20 to 50 per cent |7 (1987), v~. The 2.58 per cent premium is offset or more than offset by the wholesale discount.

As noted above, the more specialized the portfolio, the greater the price variance over time and the greater the opportunity for gain or loss. By choosing the right antique type, instead of a more diverse portfolio, one could have done better. For example, a portfolio of beds bought and held during 1967-1977 yielded 15.3 per cent (row 2, column 7) when T-bills were yielding 5.66 per cent. The largest wholesale discount of 50 per cent would still have yielded almost 3 per cent net. But, as with all relatively specialized portfolios, ab initio a sub-set like this would have been riskier, especially relative to financial instruments.

During the second half of the two decade span, opportunities for gainful investment in antiques deteriorated substantially. The antiques' overall rate of return fell from 8.24 per cent in the first decade to only 5.71 per cent in the second decade (rows 2 and 3, column 1), while T-bill rates were soaring (rows 2 and 3, column 10) from 5.66 to 8.99 per cent. Every portfolio depicted in the table, even beds, would have yielded a clear loss.

IV. Conclusion

The results in Table I are generally consistent with those of previous research. Studies of the monetary rate of return on investments in collectibles, through long-term price appreciation, typically conclude they are a poor investment relative to financial instruments.

Since collectibles, in fact, are more than mere financial instruments to their owners (who otherwise presumably would be collecting financial instruments, instead), one suspects we are measuring collectibles' rates of return against the wrong criteria. Along with pecuniary gains, we ought to be including nonpecuniary user benefits. Only in one case, that of bottled wine, does the pecuniary yield appear to reflect the natural increase in user benefit at the final imbibing.

The problem is that user benefits, from daily association with an aesthetic object, are not readily disentangled from the object's financial yield. If use and ownership were separate, the user paying a rent to the owner for exclusive daily association with the collectible and the owner getting, in addition to the rent, the financial yield from any price appreciation, then the market would reflect the total return--rent plus capital gain--from the collectible. One suspects the mean value of this total return would easily stand comparison with yields on financial instruments.

Such rental arrangements seem very rare. At least, we have no public data base of the kind. There is the problem of agency. The principal hindrance to such a market may be the lack of motivation in the renter to preserve and protect the aesthetic object as would an owner-in-possession. Buying collectibles to rent out, therefore, may simply be unprofitable.

Paul Graeser Northern Illinois University DeKalb, Illinois

1. The benefit from consuming a great vintage wine is a user benefit. The wine's value as a collectible inheres in its age and vineyard, which are aesthetic qualities.

2. In six per cent of the sub-type cells, |K.sub.ijt~ = 0. The fortran program written for data management deals with these empty cells, thus avoiding the obvious bias. (I am indebted to Susan Porter-Hudak for this useful program.)

3. The method is similar to computing the well known Consumer Price Index. One difference is that the antiques index uses the entire 20 year period to determine the weights, instead of the less reliable Laspeyre's single base-year weights. Also, the antique "product" itself does not depreciate. Indeed, the antiques market is essentially a circulating stock: vendor offerings of antiques, unlike offerings of typical consumer goods, collectively change little or none over time.

References

1. Anderson, Robert C., "Paintings as an Investment." Economic Inquiry, March 1974, 13-26.

2. Baumol, William J., "Unnatural Value: or Art Investment as a Floating Crap Game." American Economic Review, May 1986, 10-14.

3. Federal Reserve Bulletin, 1970-1988. Washington, D.C.: Board of Governors of the Federal Reserve System.

4. Frey, Bruno S. and Werner W. Pommerehne, "Is Art Such a Good Investment?" The Public Interest, Spring 1988, 79-86.

5. Jaeger, Elizabeth, "To Save or Savor: The Rate of Return to Storing Wine." Journal of Political Economy, No. 3, 1981, 584-92.

6. Keen, Geraldine. Money and Art: A Study Based on the Times-Southeby Index. New York: G.P. Putnam's Sons, 1971.

7. Kovel, Ralph and Terry Kovel. Kovels' Antiques and Collectibles Price List. New York: Crown Publishers, annually 1968-1987.

8. Krasker, William S., "The Rate of Return to Storing Wines." Journal of Political Economy, No. 6, 1979, 1363-67.

9. Ross, Myron H. and Scott Zondervan, "Capital Gains and the Rate of Return on a Stradivarius." Economic Inquiry, July 1989, 529-40.

10. Stein, John P., "The Monetary Appreciation of Paintings." Journal of Political Economy, October 1977, 1021-35.

Printer friendly Cite/link Email Feedback | |

Title Annotation: | Communications |
---|---|

Author: | Graeser, Paul |

Publication: | Southern Economic Journal |

Date: | Apr 1, 1993 |

Words: | 2075 |

Previous Article: | Modern Chinese paintings: an investment alternative? |

Next Article: | Two views of applied welfare analysis: comment. |

Topics: |