Art as a financial asset : a performance analysis of art as an alternative investment

(1)

Art as a Financial Asset

A Performance Analysis of Art as an Alternative Investment

Bachelor Thesis

Economics and Finance

By David Kok

10745068

(2)

David Kok - University of Amsterdam 2

Statement of Originality

This document is written by David Kok, who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(4)

1. Introduction

On 15 November 2017, the painting Salvator Mundi of Renaissance master Leonardo da Vinci was sold for an all-time record of 450,312,500 USD at auction house Christie’s (Gerlis, 2017). This all-time record combined with a recently published report called ‘The Art Market 2018’ (McAndrew, 2018) sheds a light on the current state of the art market: it is booming.

According to the writer of The Art Market 2018 and founder of Arts Economics Clare McAndrew, global sales surged from 56.6 billion USD in 2016 to 63.7 billion USD in 2017. Because the total value of the global art market diluted 16% since 2014 until the end of 2016, the increase of 12% in this specific year is remarkable.

The volume of sales grew at 8% in 2017. These volumes, measured in total transactions, were at the highest level since the start of the financial crisis in 2008 (McAndrew, 2018). The difference between the numbers mentioned above can be attributed to increasing interest which is reflected in average market prices (Artprice, 2018). To illustrate this, the Artprice.com index shows an increase of 36% in average market price for art since the start of this century.

From the above there can be concluded that the art market is becoming more attractive to investors. There is a widespread interest in buying art and the figures show that large amounts of money are involved (Artprice, 2018). According to The Wealth Report (Knight Frank Research, 2018), 35% of the High Net Worth Individuals (HNWI) is investing in the market for art and collectibles and this percentage is expected to increase for at least the following ten years, respectively.

According to Renneboog and Spaenjers (2013), high-income consumer confidence is the main determinant for predicting art prices. Art is seen as a collector good’, in the way that it has features of both speculative and durable consumer goods (Stein, 1977). The demand for art increases when income rises proportionally. Also, Goetzmann (1993) found evidence on a positive correlation between art and wealth. This implies that when income increases, the demand for art will also increase. Therefore, it is reasonable to think that the art market will continue to grow the coming years.

In the past few years, several art funds have been established to guide investors in the art market, like ANTHEA and The Fine Art Group (Campbell, 2008). These companies provide assistance to investors who prefer to invest in art purely for financial gain. A survey by McAndrew (2018) shows that 32% of collectors say the return on their art investment is important. When the same survey was conducted on the HNWI population, this percentage was even larger: 47 (McAndrew, 2018). Due to the origination of such art funds, investors

(5)

David Kok - University of Amsterdam 5 now have the possibility to indirectly invest in art through investing in diversified art portfolios with the aim of only achieving financial gains. In addition, the availability of price indices, data and art market reports have increased the willingness to invest in the art market as found by Campbell (2008).

In addition, Campbell (2008) found a relationship between relatively bad performance and the use of investment opportunities. She states that relatively bad performance of traditional assets leads to an increasing use of alternative investment opportunities like hedge funds, real estate and emotional assets like art. This deviation in the use of certain investment opportunities can be attributed to the investors’ need of portfolio diversification. During this thesis, only the use of art as an alternative investment opportunity will be discussed.

This thesis investigates the use of art as an alternative investment. First of all, the diversification possibilities of art will be discussed on the basis of a correlation matrix. Then, a risk and return analysis will be done. Eventually, from Markowitz portfolio theory two optimal portfolio allocations will be constructed. By the use of the Sharpe index and the minimum variance portfolio a performance analysis is conducted. This analysis considers a portfolio excluding art and a portfolio including art as a financial asset. The main question in this thesis is to what extent art can be seen as a profitable alternative investment product compared with traditional financial asset classes. In this paper, these assets are referred to as stocks, government and corporate bonds, commodities and real estate.

To study this, the remainder of this paper is structured as follows. In the first chapter, an overview of the previous literature on this topic will be given chronologically to provide a clear academic framework. In the second chapter, the data collection in discussed. This is divided into three subparagraphs: concerning the art market dataset; the financial market dataset; and the limitations of the datasets used. Further, in the third chapter, all elements of the methodology that is used to conduct this research will be explained. In the next chapter, the results will be presented and interpreted. After that, the last chapter consists of the conclusions that can be drawn from these results and a brief summary of the paper will be given.

(6)

2. Literature Review

In the academic literature so far, several publications are shown in the field of art and finance. In this paper, the most important publications on this topic will be discussed chronologically to get a hold on the development in this specific area of finance. For each article, the most relevant components are explained to provide a clear academic framework. These results will be of influence on answering the research question with respect to the performance of art as an alternative investment product. First of all, the previously examined performances of art as a financial asset will be discussed by describing the key findings in literature. Subsequently, articles regarding portfolio diversification are described. Eventually, methods and limitations concerning the art price indices are explained.

2.1 Art as an Alternative Investment

The research on financial rates of return on art works started with two important publications. To begin with, Anderson (1974) published one of the first articles on this topic. Firstly, he looks at the variables affecting art prices and concludes that there are only two observable variables significantly influencing the art market prices: the reputation and the artistic ‘delivery’ of the artist. In his paper, Anderson states that the art market demand can be divided into three categories, museums; dealers; and collectors determine the total demand in the market. When it comes to art as an investment product, only private collectors are concerned. Further, the question arises in what respect art can be distinguished from consumption goods. Therefore, it is important to structure a few terms. It is described that there are two types of services: financial –and consumption services. Art as an investment provides both. Consumption services consist of decorative and aesthetic- prestige services. From 200 years of data up till 1960 it is stated that the turnover rate on art works is extremely low compared to other financial assets. So, the consumption value of art can be seen as the main motivation to purchase art works. This is confirmed by analyzing the results that show that art is considered to be a riskier investment and is associated with a lower rate of return in the long term.

Three years later, Stein (1977) confirms that paintings are both consumer durables and capital assets. In addition to the model, he incorporated the variable ‘viewing pleasure’ to show the return on aesthetic-prestige services together with the financial return. For this, Stein uses the capital asset pricing model for durable goods. Then, the Jensen’s Theory is used to “evaluate the financial behavior of individual assets and portfolios” (Stein, 1977, pp.1028). He concludes that paintings are ‘inefficient’. They are regular assets, given that investors are

(7)

David Kok - University of Amsterdam 7 interested in the financial benefits of their investment and value the variable ‘viewing pleasure’ at 1.6 percent per year. In short, investors should only invest in art works when the aesthetic value they give to their investment is big enough to compensate for the financial benefits of investing in traditional assets, like stocks. In this, Stein agrees with Anderson (1974). According to Stein (1977, pp. 1031) this is due to the high degree of nonsystematic risk in the returns of art works. It is not profitable to invest in art purely for financial purposes.

Ten years later, Baumol (1986) writes a critical report on art market investments. They are referred to as a ‘floating crap game’. In other words, determining the value of art is a completely ‘unnatural’ process. Namely, prices of art float randomly over time and cannot be predicted with any certainty. Also, Baumol says no financial gain can be derived from investing in art. Even more transparency of information about prices and features in the art market will not lead to more positive results regarding the performance of art as an investment product. In his paper, Baumol uses a dataset of deceased artists from 1652 – 1961 to compare the financial returns of art with interest rates on long term bonds. Briefly, the average real return on art is 0.55% and the return on bonds is 2.5%. Investors will not be able to meet the opportunity cost on their investment. Baumol points out that the aesthetic pleasure should be the primary motivation to invest in art.

In the 1980’s, the art market is booming. In this period, ‘art as an investment’ is becoming more and more popular in the United States and Europe. The question arises why it is possible that Anderson (1974), Stein (1977) and Baumol (1986) claim that art is not a profitable alternative investment. Frey and Pommerehne (1989) argue this. There are two major shortcomings of the above-mentioned papers. Firstly, they make use of large datasets that consist over more than 200 years of art price data. This data only goes until the 1960’s while the first twenty years after this period the most relevant information can be required. Secondly, the data is mostly based on Christie’s and Sotheby’s auction figures, and does not include any data on other European auction houses. Frey and Pommerehne try to overcome these limitations by including appropriate precautions. This leads to the following results. The financial rate of return on paintings lies much lower than the returns on traditional investments. Given that art investments are riskier than other assets, they must yield a higher return to compensate for that risk to be at least equally attractive for investors. So, there must be a consumption value as stated in previous research. From a pure financial point of view, there are no benefits associated with art investments, except for aesthetic return and the tax

(8)

David Kok - University of Amsterdam 8 advantages that art investments can provide, depending on the regulation of the country in question.

After that, Goetzmann (1993) estimates the art market returns in the period of 1715 – 1986. The results are more or less the same with respect to previous publications. The return on art is relatively low compared to the return on long term bonds. Nevertheless, for the years beyond 1850, Goetzmann finds that art prices have appreciated significantly over time and that aggregate returns on art investments are higher than returns on stocks and long-term bonds. However, Goetzmann’s research also shows that art cannot be seen as an attractive investment tool. The results point out that art prices are highly correlated to stocks. Therefore, Goetzmann assumes that the art market is positively influenced by the increase in wealth. This is the reason why, during the post-World War II period – a period of economic prosperity – the art market returns were high.

Pesando (1993) specifically looked at prints for the period of 1977 – 1992 and compares this small component of ‘art’ with traditional assets. He found that the annual real return rate of these prints is lower than the returns on stocks and bonds, but the risk of investments is equal. Therefore, he concludes that prints not a profitable investment purely for financial gains. Pesando also finds that the masterpieces in prints underperform the aggregate returns.

Further, in 1995 Frey and Eichenberger come with new insights about art market. Art should not be handled as if it is an ordinary asset or a commodity. The art market has some special features that have to be taken into account in the research model. The most important two characteristics of the market that should be included are transactions costs and behavioral differences. Firstly, transaction costs are much higher for buying art works compared to traditional investments. As mentioned earlier, dependent on the countries regulation, there can be tax advantages. Secondly, all demanders for art have different incentives to invest in art of which financial gain is one of the least important.

Another article on the return on alternative investments, is published by Mei and Moses (2002). In here, the US art market in the period of 1875 – 2000 is looked at. First of all, art returns are at 4.9% per year, higher than the fixed-income securities. If we look at previous publications, we see differences in the volatility of art prices and correlation with other assets. The volatility is less and the correlation is much lower. Still, the standard deviation of art investments is much higher than the stock market risk. Thus, common stocks outperform alternative investments. Another finding, in line with Pesando (1993), is that the expensive paintings underperform the market index. This is also called “the

(9)

David Kok - University of Amsterdam 9 underperformance of masterpieces”. To conclude, when it comes to financial purposes investors should only get into art investments to smooth the transaction costs over time.

That being the case, Renneboog and van Houtte (2002) looked at the risk-return trade-off of Belgian art (1970-1997) compared to stocks, using financial performance measures. They show via the Markowitz efficient frontier that art does not have a positive effect on the diversification possibilities of an investment portfolio. Art is outperformed by traditional assets by their high riskiness and transaction fees. Art returns significantly underperform stock returns, with comparatively low Sharpe ratios associated to the portfolios. At last, Renneboog and van Houtte describe the illiquidity of the art market and the extra costs that reduce art returns. These are transaction costs, yearly insurance premiums and payable resale rights.

More recently, Campbell (2008) wrote a paper on art as a financial asset. She discusses the current state of the art market. There is growing potential in art investments and more and more funds are established to contribute to this. Directly investing in art is risky, but it yields returns which involve certain elements of viewing pleasure that can compensate for poor financial performances. On the other hand, indirectly investing in art does not yield anu aesthetic pleasure, but it has high potential. The art market is still in its infancy when it comes to investment opportunities and financial gain. Campbell describes art funds that assist investors who’s only concern is financially benefiting from their investment. According to Campbell there are two ways of indirectly investing in the market for art and collectibles, via alternative investment vehicles (AIV) or an art mutual fund (AMF). Subsequently, she performs a test on whether art investments yield returns comparable to those of government and corporate bonds, real estate, commodities, and stocks. She finds that there is low correlation between art and these assets, which means there should be high diversification potential. However, Campbell’s results show a significantly high degree of nonsystematic risk in art returns. There is high volatility in the art market. Also, the market is very illiquid compared to the equity market. Campbell values the aesthetic return on art investments at 1,6 percent per year. Considering this, she concludes that art is no more or less attractive as an alternative investment than the traditional financial market assets. Art has to be considered as an asset class, but does not yield outstanding performances compared to other assets. She draws the same conclusions as her predecessors did, but adds to it that art has portfolio diversification benefits, due to the significantly low correlation with other assets. Furthermore, with art funds showing annually increasing results, the expectations of the profitability of art investments is hopeful.

(10)

David Kok - University of Amsterdam 10 At last, Renneboog and Spaenjers (2013) published an academic paper called ‘Buying Beauty’. In the time this article was produced the art market was becoming more and more transparent. The possibilities of getting large datasets and art market information are growing yearly. Renneboog and Spaenjers look at “the price determinants and investment performances of art” (Renneboog and Spaenjers, 2013, p. 36). Consequently, more recent data is used: registered auction prices from 1957 to 2007. This is a more relevant representation of the current art market trends. During this period art prices have been appreciated approximately 4% per year. Since 1982 the real return of art is at 5.19% per year. From performance analysis is shown that art is outperforming bonds in this period. Bond prices show similar movement over time, but operate at much higher risk. Due to this finding, the academic relevance of previous and coming studies can be easily substantiated. Renneboog and Spaenjers mention three main reasons for the increasing demand of investors in the art market. Firstly, the amount of multi-million-dollar sales is growing substantially. Secondly, the number of HNWI’s is increasing yearly. Finally, just as Campbell (2008) showed, the demand for more portfolio diversification is also rising. In the paper, a comparison is made between art and other financial assets. Resulting in art outperforming the following assets: gold, real estate and commodities. This is in contrast to Campbell’s (2008) research. But, the overall performance of art as an investment, regarding the risk-return trade-off, is less positive than any other asset class. Lastly, Renneboog and Spaenjers also point out that they did not find any prove to confirm the underperformance of masterpieces as suggested by Pesando (1993) and Mei and Moses (2002).

2.2 Portfolio Diversification

In this paragraph, literature about art as a profitable asset class for portfolio diversification will be discussed. Hereby, three corresponding publications are used chronologically.

With this intention, it is important to give the term ‘portfolio diversification’ some more explanation. Diversification implies a mechanism to minimize risk and yield higher returns by investing in a portfolio consisting of different asset classes (Bodie, Kane and Marcus, 2011). Mei and Moses (2002) say a portfolio is better diversified when the sample used for testing is larger. This is why they used a larger sample and this resulted in their art index being less volatile than other assets and showing low correlation with other assets classes. They belief art is “appropriate” (Mei and Moses, 2002, p. 1666) as a diversification tool in the long-term.

(11)

David Kok - University of Amsterdam 11 Two years later, Worthington and Higgs (2004) investigate the diversification possibilities of art in portfolio management. They find that art is riskier and yield lower returns than traditional assets like corporate bonds and United States equity. They also find that there are no diversification benefits attached to including art in the asset portfolio. Although a portfolio consisting solely of art can be considered as profitable. Also, low correlation between art markets could reduce the portfolio’s risk causing diversification benefits. Another finding is that by dividing the art index into several art categories like Campbell (2008) did in her study, also some diversification is accomplished. To summarize, Worthington and Higgs (2004) show that art has diversification benefits due to the low correlation with other assets. On the other hand, the risk and return performance of art compared to traditional assets is such that they would not recommend investing in the art market. Renneboog and van Houtte (2002) agree with this. Just like Worthington and Higgs they use the Markowitz theory to construct an efficient frontier. This frontier of botch art and equity investments does not shift upwards, which means that there is no substantial diversification potential in this portfolio.

Finally, Campbell (2008) explains why it is useful to check the correlation between art and other assets. “The level to which these assets can reduce risk in an asset portfolio depends crucially on the extent to which the returns are correlated with each other.” (Campbell, 2008, p. 74). Thus, portfolio improvement is about reducing the risk, while maintaining the returns. To determine the optimal portfolio allocation Campbell (2008) claims an assumption has to be made. The historical distribution of returns has to be used to predict future returns. This is why a long-time horizon is appropriate. Campbell concludes that there are diversification opportunities in the art market.

2.3 Art Price Indices

In this section, the methods of computing art price indices are explained in connection with the articles discussed in the previous paragraphs. According to Kraeussl and Lee (2010), there are three methods for constructing an art price index: the repeated-sales regression; the average prices model; and the hedonic regression. Several other methods for computing price indices exist, but in this paper, due to irrelevancy of other methods, only these three methods will be treated.

(12)

David Kok - University of Amsterdam 12 2.3.1 Average Prices Method

This method is used by Renneboog and van Houtte (2002) and is called a ‘regular sampling method’. They use the mean and median of the auction prices as an estimate for the overall mean value of the art works with data from 1946 to 1968. On the other hand, Stein (1977) uses the geometric mean as a proxy for the mean value of the art price index. This is comparable to the method that is described as a ‘easy’ method for constructing price indices by Kraeussl and Lee (2010). They use average and median art prices to compute the so-called ‘naïve art price index’. An assumption that is done by Renneboog and van Houtte (2002) in addition to this method is that the quality of art works has to be smoothed over time. So, the average quality of art traded on the market is constant. This method is commonly used for constructing the consumer price index. This can be explained on the basis of Stein’s (1977) study. He used a dataset from 1946 to 1968 as a representative for current and future prices. So, when art has to be valuated that is not sold in this particular period (not in the dataset), another work of the same artist or with the same features can be used as a close substitute for pricing the art work. This leads to a disadvantage mentioned by Renneboog and van Houtte (2002): the valuation of an art work is depending on the subjectivity of choosing a substitute.

2.3.2 Repeated-Sales Regression Method

Thereafter, the repeated-sales regression (RSR) is discussed. As Table 1 shows, this is the most commonly used method for constructing an art price index. Kraeusll and Lee (2010) define this as a method that estimates average return on the art works in the chosen period of time. This is with respect to art that has been sold twice before. Baumol (1986) skips all the art works that have not been resold within 20 years from their previous sale. In this way, Baumol tries to avoid speculative sales. This is due to the main disadvantage of this approach: only art works that have been sold repeatedly are included in the dataset. This causes a sample selection bias and will lead to a dataset that is not representative for the entire population (Kraeussl and Lee, 2010).

2.3.3 Hedonic Price Method

The relevance of the following model is discussed by Chanel, Gérard-Varet and Ginsburgh (1996). In their opinion, the hedonic pricing model is used in the case of heterogeneous products with infrequent trading. Renneboog and Spaenjers (2013) use a hedonic regression on a dataset of more than one million art works. The main advantage of this approach is that all information and all observation can be tested. This method does not only take into account

(13)

David Kok - University of Amsterdam 13 the resale figures of auction houses, but involves all auction house transactions over the years (Chanel et al., 1996). The conclusion is that, when applied to a large dataset, the hedonic approach will provide a more useful basis for estimating future returns and the efficiency of the art market.

Table 1

Estimated Fine Art Performance, 1716-2007

This is constructed via Campbell (2008, p. 67) and updated until Renneboog and Spaenjers (2013).

Author Year Sample Period Method Nominal Return Real Return Standard Deviation

Anderson 1974 Paintings in general 1780-1960 1780-1970 Hedonic RSR 3.30% 3.70% 2.60% 3.00% Stein 1977 Paintings in general 1946-1968 Average prices 10.50%

Baumol 1986 Paintings in general 1652-1961 RSR 0.60%

Frey and Pommerehne 1989 Paintings in general

1635-1949 1653-1987 1950-1987 RSR 1.40% 1.50% 1.70% 5.00%

Goetzmann 1993 Paintings in general

1716-1986 1850-1986 1900-1986 RSR 3.20% 6.20% 17.50% 2.00% 3.80% 13.3% 5.65% 6.50% 5.19%

Pesando 1993 Modern prints 1977-1992 RSR 1.51% 19.94%

Chanel et al. 1996 Paintings in general 1855-1969 1855-1969

Hedonic RSR

4.90% 5.00%

Mei and Moses 2002 American. Impressionist, and Old Masters

1875-1999 1900-1986 1900-1999 1950-1999 1977-1991 RSR 4.90% 5.20% 5.20% 8.20% 7.80% 4.28% 3.72% 3.55% 2.13% 2.11% Renneboog and Van Houtte 2002 Belgian art 1970-1997 Average prices 5.60%

Campbell 2008 Paintings in general 1976-2002 Average prices

RSR 6.56% 8.08%

(14)

3. Data

3.1 Art Market Data

As literature shows, it is possible to get large datasets on auction results. As previously mentioned, there is more and more information transparency in the art market the last decades (McAndrew, 2018). Commercial databases have been established during this period to provide more information on the art market. One of them is Artprice, the world leader in art market information. The dataset used in for this research is derived from the ‘2017 Report on the Art Market’ (Artprice, 2018). This is an art market report that gives the most recent trends and figures on the art market.

The dataset is an art price index that is constructed using the repeated-sales method. It contains of quarterly annual data from 01/01/1998 until 01/01/2018 – a 20-year time-span. This sample involves 80 observations (see Appendix for the full dataset that is used in this thesis). From Renneboog and Spaenjers (2013) we know this should be a sufficient sample. More information on this price index is found in ‘2017 Report on the Art Market’ (Artprice, 2018). Namely, the prices stated in this set indicate public auction results that include the buyer’s premium. The index concentrates only on fine arts. This means that the dataset contains of the following: paintings; photographs; drawings; prints; videos; tapestries; installations; and does not obtain antiques; furniture or anonymous cultural goods (Artprice, 2018).

Thereby, the index categorizes art into five subgroups. Firstly, ‘old masters’ refers to works by artists born before 1760. Secondly, ‘19th_{century’ refers to artworks created by}

artists between 1760 and 1860. Thirdly, art works of artists born between 1860 and 1920 are categorized as ‘modern art’. Fourthly, ‘post-war art’ represents the period of 1920 to 1945. Lastly, ‘contemporary art’ is used to define the works by artists born after 1945.

Campbell (2008) criticizes the current art market databases and points out their limitations. To begin with, the art market data is mostly based on the figures of the main auction houses. The auction results only do not give a clear inside into the real performance of the art market, because they cover only a slight aspect of the entire market movements. Besides that, The Art Market 2018 (McAndrew, 2018) shows that the aggregate sales by dealers account for a larger share of the total market sales compared to auction sales, 53% against 47%. This is why Campbell (2008) states that it is a big problem for art market research that the dealer market is often ignored. This is due to unobtainable data on dealer transactions. With taking the dealers into account, Campbell (2008) finds that the return on investment of art would reduce. She states that it is likely that dealers will buy at lower prices

(15)

David Kok - University of Amsterdam 15 and sell at higher transaction costs (Campbell, 2008). Besides that, also the art fund sales data is not publicly accessible.

Another limitation of the art market dataset is the use of the repeated-sales regression to construct the art price index. As mentioned before, the repeated-sales method is a method by which solely art works that are sold at least twice at auction are taken into account. This means the dataset is an incomplete representation of the art works sold in the market. This method causes a sample selection bias (Kraeussl and Lee, 2010).

3.2 Financial Market Data

The financial data in this study is obtained via Datastream and Quandl. These financial market datasets consist of 80 observations of quarterly annual data from 01/01/1998 to 01/01/2018, similar to the art market data. The data collection purely focuses on showing a fine representation of the financial asset market of the past twenty years. Hence, in this paper is chosen to use data of the global stock market; government and corporate bonds; commodities; US treasury bills; and real estate as proxies. Moreover, the total return indices of these asset classes will be used, because the variables will be compared with art returns in a performance analysis. Thereby, all price data used in this paper is given in US dollars and returns are shown in percentages.

In this paper, the MSCI World Equity index is used as a benchmark for the stock market, the S&P GSCI Commodity index represents the commodities and for the real estate market the MSCI World Real Estate index is used. Further, for finding a representation for the government bonds, the US and UK 10-year government bond index is used. This is because the United States and the United Kingdom are the first and third largest art markets in the world, accounting for 42% and 20% of sales by value respectively (McAndrew, 2018). Therewith, the Citigroup USBIG Treasury index is used as a benchmark for US short term treasury bills to use as a proxy for the risk-free rate. Lastly, the Merrill Lynch US corporate bonds index is used for representing the corporate bond market of the United States.

A limitation of this, is that China is not involved in the dataset. That would have led to a completer dataset in comparison to the art market data. Whereas China is the second largest art market in the world, taken into account 21% of sales by value (McAndrew, 2018).

Another limitation is, this data concerns total returns instead of prices. Via the possible use of different constructing methods, the eventual data could be incomparable.

(16)

4. Methodology

In this section, all methods and theories will be explained. This research is done on the basis of the portfolio theory of Markowitz and the performance measurement of Sharpe. The relevant formulas and elements of these theories are constructed using academic literature.

First of all, from the art price dataset the art returns of the market and the returns of the art schools, i, are computed by continuously compounding returns (1). The returns are given as the natural logarithmic values of the art prices at time, t, such that the change in prices, 𝑝𝑖,𝑡,

denotes the rate of change of prices, 𝑝_𝑖,𝑡 (Campbell, 2008):

(1) ∆𝑃𝑖𝑡 = 𝑙𝑛 ( 𝑝𝑖,𝑡

𝑝𝑖,𝑡−1) × 100

This is how the quarterly returns are computed. From this, the geometric returns are calculated. As derived from Renneboog and Spaenjers (2013), these are the averages of the continuously compounded returns over the period 01/01/1998 – 01/01/2018. In this paper, these averages are called the average returns. To calculate the returns of the traditional assets the arithmetic average rate of return is used. This is because the financial market data shows the total return indices – instead of the price indices – of the traditional assets (Worthington and Higgs, 2004). The formula to compute the arithmetic returns on time, t, is stated below (2). After this, the development of market returns of the art and traditional investments over the period 1998 – 2018 is graphed.

(2) ∆𝑅_𝑡= (𝑅𝑡− 𝑅𝑡−1

𝑅𝑡−1 ) × 100

Then, the datasets are merged and the descriptive statistics are derived from the merged data. From this, the volatility (3) of the quarterly returns is computed by using the following formula (Campbell, 2008):

(3) 𝑉𝑜𝑙𝑎𝑡𝑖𝑙𝑖𝑡𝑦 = 𝑆𝑡𝑑.𝐷𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡ℎ𝑒 𝐴𝑠𝑠𝑒𝑡

√𝑁𝑢𝑚𝑏𝑒𝑟𝑜𝑓 𝑃𝑒𝑟𝑖𝑜𝑑𝑠

The standard deviation can be derived from the descriptive statistics obtained via Stata SE 15. To calculate the quarterly volatility of asset returns the number of periods in the denominator must be 20. This is because the dataset consists of 80 observations in total. So, the standard

(17)

David Kok - University of Amsterdam 17 deviation is divided by the square root of 20 to get the volatility statistics. After this, the Sharpe ratio for all variables is conducted. The Sharpe index (4) is used to show the average excess return relative to the riskiness of the fund and is acts as a performance measurement tool for investment portfolios (Bodie et al., 2011). In addition, Bodie et al. (2011) state that the Sharpe ratio is an adequate method to measure the risk to return trade-off of a diversified portfolio. It is not adequate when used for individual assets.

(4) 𝑆ℎ𝑎𝑟𝑝𝑒 𝑖𝑛𝑑𝑒𝑥 = 𝑅𝑃−𝑅𝑓

𝜎𝑃

The Sharpe ratio of each asset portfolio is calculated by subtracting the risk-free rate from the expected portfolio return and divide this by the risk (standard deviation) of the portfolio (Bodie et al., 2011). Stein (1977) uses short term US treasury bills to serve as a proxy for the risk-free rate. He states that no particular biases occur, when using this as an estimate for a riskless asset. In this paper, also the short-term US treasury rate will be used for an approximation of the risk-free rate.

Now the quarterly volatility and return of each portfolio are known, the return to risk ratio (5) can be computed. This is the level of return that is derived per unit of risk (Campbell, 2008). Hereby, risk is measured using the volatility of returns. Subsequently, a risk-return trade-off is graphed.

(5) 𝑅𝑒𝑡𝑢𝑟𝑛 𝑡𝑜 𝑅𝑖𝑠𝑘 𝑅𝑎𝑡𝑖𝑜 =𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑅𝑒𝑡𝑢𝑟𝑛

𝑉𝑜𝑙𝑎𝑡𝑖𝑙𝑖𝑡𝑦

Further, an important part of showing the diversification possibilities of art is by presenting the correlations across sectors and across art schools. This is given by a correlation matrix (Worthington and Higgs, 2004). A covariance –and variance matrix shows the levels of risk between two variables from the dataset (Markowitz, 1952).

Eventually, with the help of the Markowitz portfolio theory, the optimal portfolio allocations will be derived. Investments should be analyzed by how they affect the total risk and return of the portfolio, rather than being evaluated solely on the basis of their own performance (Markowitz, 1952). In this paper is looked at the minimum variance – portfolio. As described by Markowitz (1952), this is a portfolio that consists of several individually risky assets, but combined into a diversified portfolio the risk is minimized. To calculate the minimum variance portfolio the covariance –and variance matrix is needed. Then, the weights

(18)

David Kok - University of Amsterdam 18 of the assets in portfolio are attributed (Markowitz, 1952). When this is derived, the optimal return and variance of the portfolio can be calculated.

The following formulas are derived from Bodie et al. (2011). Suppose there are N risky assets, in this study N = 13. Then, R denotes the return on the asset, respectively (6). Then the return on the portfolio is given as follows (6):

(6) 𝑅_𝑃 = ∑𝑁_𝑛=1𝑤_𝑛𝑅_𝑛

In here, w denotes the weight of investment n in the portfolio (7). Notice that:

(7) ∑𝑁 𝑤_𝑛

𝑛=1 = 1

From this, the portfolio return (8) and variance (9) can be computed.

(8) 𝜇𝑃 = 𝐸(𝑟𝑝) = ∑𝑛𝑖=1𝑤𝑖𝐸(𝑟𝑖)

(9) 𝜎_𝑝2 = ∑_𝑖=1𝑛 ∑𝑛_𝑗=1𝑤_𝑖𝑤_𝑗𝐶𝑜𝑣(𝑟_𝑖, 𝑟_𝑗)

After this, the efficient frontier of risky assets is calculated. This is “the identification of the efficient set of portfolio’s” in a graph (Bodie et al., 2011, p. 222). According to the Markowitz portfolio theory the efficient frontier shows the expectation of a set of portfolios, given a certain level of risk, that yield the highest returns (Markowitz, 1952).

Rational investors choose a portfolio on this frontier. Suppose, M is the covariance –and variance matrix and f being a constant that is equal to the risk-free rate. Then, the weights that would optimize the diversified portfolio are calculated using this formula (10):

(10) 𝑤_𝑖 = 𝑀−1(𝜇𝑃−𝑓)

∑ 𝑀−1_(𝜇 𝑃−𝑓) 𝑁

𝑛=1

Now the weights of the minimum variance portfolio are calculated, the efficient frontier is constructed by adjusting the returns for the same portfolio and see how the portfolio variance reacts to this.

Eventually, the Sharpe ratio of the weight-adjusted optimal portfolio allocations is computed. Then, the analysis of the performance of these portfolio’s is done. In this paper,

(19)

David Kok - University of Amsterdam 19 five different risky portfolios are used for this analysis: a portfolio excluding art; a portfolio including the global art index; a portfolio including the various art schools; a portfolio including the global art index, excl. US treasury bills; and a portfolio including the various art schools, excl. US treasury bills. The efficient frontier of the optimal portfolio allocation including the global art index is shown in a graph.

(20)

David Kok - University of Amsterdam 20 0 50 100 150 200 250 300 350 1/1 /199 8 11 /1/19 98 9/1 /199 9 7/1 /200 0 5/1 /200 1 3/1 /200 2 1/1 /200 3 11 /1/20 03 9/1 /200 4 7/ 1 /2 00 5 5/1 /200 6 3/1 /200 7 1/1 /200 8 11 /1/20 08 9/1 /200 9 7/1 /201 0 5/1 /201 1 3/1 /201 2 1/1 /201 3 11 /1/20 13 9/ 1 /2 01 4 7/1 /201 5 5/1 /201 6 3/1 /201 7 1/1 /201 8 Old Masters 19th Century Modern Post-War Contemporary

5. Results

In this chapter, the results of the research done as described in previous sections will be carried out. On the basis of the following figures and tables the results will be discussed.

First of all, the price development of art price for the period 1998 – 2018 is shown per art school category (Figure 1). By this, it is shown that art prices increase enormously during the financial crisis of 2008. After that, prices decreased to the point where the post-war and contemporary art prices are the highest. Only 19th century and old master art works have decreased in prices.

Figure 1

Art price development, 1998-2018.

The return statistics of the merged datasets are stated in Table 2. When it comes to art, post-war and contemporary art yield the highest returns, but none of the art schools nor the global art index yields a higher return than any of the financial market assets, except for commodities. Commodities, art in general and real estate have the largest standard deviation, which automatically leads to the highest volatility and therefore risk. Also, does art in general yield a lower average rate of return than all the other asset classes. When we look at the Sharpe ratio of each individual asset, we see a positive value for global stocks, US and UK government bonds, real estate and US corporate bonds. Again, commodities and art underperform.

(21)

David Kok - University of Amsterdam 21 Table 2

Return statistics of art and financial market data from 1998 – 2018.

For the development of the returns in the art –and financial market Figure 2 is given. In the period of the financial crisis both art and traditional assets show large fluctuations. Especially for post-war art, commodities and art in general the change of return over time is volatile compared with the other assets. The art market, in general, is more volatile than the financial market.

Figure 2

Development of returns of both art and financial assets, 1998-2018.

Variable Observations Mean

Standard

Deviation Minimum Maximum

Quarterly

Volatility Sharpe Ratio

Global Art Index 80 0.5216555 10.68623 -21.26095 33.17448 1.194756835 -0.059248818

Old Masters 80 -0.2384616 6.960248 -16.70199 21.83719 0.778179383 -0.200174419 19th Century 80 -0.3464606 5.111997 -14.56712 13.71578 0.57153864 -0.29367439 Modern 80 0.2157527 3.725459 -12.69242 8.026324 0.416518979 -0.252062712 Post-War 80 0.8734712 4.881326 -11.34484 15.74555 0.545748838 -0.057634094 Contemporary 80 0.5705105 6.784169 -15.74586 15.9032 0.758493153 -0.086125729 Global Stocks 80 1.973425 8.663486 -21.78213 21.22309 0.968607181 0.094491178 Commodities 80 0.6579216 12.81356 -46.22017 30.84221 1.43259956 -0.038777701 US Government Bonds 80 1.30502 4.179548 -6.539308 15.34127 0.467287672 0.035941207 UK Government Bonds 80 1.606939 3.410786 -4.953895 11.52025 0.381337468 0.13256094 Real Estate 80 2.256564 10.41167 -29.69807 35.70697 1.164060094 0.105819912 US Treasury Bills 80 1.154802 2.409149 -3.785866 8.417945 0.269351047 0 US Corporate Bonds 80 1.411356 2.567928 -7.490793 10.81733 0.287103078 0.099907007 Return Statistics -60 -50 -40 -30 -20 -10 0 10 20 30 40 Q1 19 98 Q1 19 99 Q1 20 00 Q1 20 01 Q1 20 02 Q1 20 03 Q1 20 04 Q1 20 05 Q1 20 06 Q1 20 07 Q1 20 08 Q1 20 09 Q1 20 10 Q1 20 11 Q1 20 12 Q1 20 13 Q1 20 14 Q1 20 15 Q1 20 16 Q1 20 17 R e tu rn s (% ) Time Return Development

Global Art Index Old Masters 19th Century Modern Post-War Contemporary Global Stock Commodities US Government Bonds UK Government Bonds

(22)

David Kok - University of Amsterdam 22 Subsequently, the correlation matrix is shown in Table 3. First of all, the global art market correlates with global stock, commodities, real estate and US corporate bonds. From all art schools, only contemporary art correlates positively with corporate bonds. Further, all schools correlate with stocks, commodities and real estate. There is a negative correlation of 30.37% and 34.14% with US and UK government bonds, respectively. Together with corporate bonds and treasury bills these give low correlations with art. With respect to the diversification possibilities of art, the low correlation between these assets and art should cause diversification benefits.

Table 3

The correlations across all variables.

Table 4 shows that global art yields 0.22 return per unit of risk. That is, compared with other assets, the lowest. The volatility of art is 2.39 and the average return is 0.52. If this is compared to the other assets: art is outperformed by all of them. If we look at the various schools, we see that modern, post-war and contemporary outperform commodities, but no more. Old masters and 19th century art works even yield negative returns per unit of risk. The return/risk ratio for art it completely outperformed by the financial market assets. Especially the assets that had low or even negative correlation with art perform well. Contemporary art correlates positively with stocks, commodities and US corporate bonds.

Global Art Index Old Masters 19th Century Modern Post-War Contem porary Global Stock Commo dities US Govern ment Bonds UK Govern ment Bonds Real Estate US Treasury Bills Us Corpora te Bonds Global Art Index 1

Old Masters 0.1559 1 19th Century 0.3328 0.204 1 Modern 0.406 0.2803 0.6369 1 Post-War 0.397 0.1313 0.4876 0.6808 1 Contemporary 0.3168 -0.0748 0.44 0.5825 0.6614 1 Global Stock 0.1276 0.2794 0.3017 0.4473 0.2759 0.1929 1 Commodities 0.2136 0.2824 0.3722 0.3801 0.3029 0.3056 0.3077 1 US Government Bonds -0.3037 -0.174 -0.1979 -0.3759 -0.2014 -0.1496 -0.6139 -0.3889 1 UK Government Bonds -0.3414 -0.1224 -0.268 -0.4185 -0.2625 -0.1796 -0.4694 -0.502 0.8646 1 Real Estate 0.1259 0.191 0.194 0.4043 0.2825 0.0993 0.7968 0.2476 -0.3861 -0.3042 1 US Treasury Bills -0.2582 -0.1492 -0.1397 -0.3241 -0.1655 -0.116 -0.6075 -0.3491 0.9736 0.8323 -0.3713 1 US Corporate Bonds 0.0072 -0.0558 -0.1169 -0.0364 -0.1003 0.0978 -0.0957 -0.0426 -0.0503 0.0023 -0.0738 -0.0311 1 Correlation Matrix

(23)

Return per Unit of Risk for each Variable

In Figure 3 the risk-return trade-off is shown. The global art market investment is the riskiest investment. It also yields the lowest return compared with the traditional asset classes. Art is outperformed by all financial assets when it comes to the return to risk-return trade-off. When it is divided into the various art schools, the results show even lower return for old masters, 19th century and modern art works. Only contemporary and post-war art yield a higher return, given that volatility is also lower. Post-war art even outperforms commodities with respect to risk and return.

Figure 3 Risk-Return Trade-Off Global Art Index Old Masters 19th Century Modern Post-War Contem porary Global Stock Commo dities US Govern ment Bonds UK Govern ment Bonds Real Estate US Treasur y Bills Us Corpora te Bonds Volatily 2.38951 0.77818 0.57154 0.41652 0.54575 0.75849 0.96861 1.4326 0.46729 0.38134 1.16406 0.26935 0.2871 Average Return 0.52166 -0.2385 -0.3465 0.21575 0.87347 0.57051 1.97343 0.65792 1.30502 1.60694 2.25656 1.1548 1.41136 Return/Risk 0.21831 -0.3064 -0.6062 0.51799 1.6005 0.75216 2.03738 0.45925 2.79276 4.21396 1.93853 4.28735 4.91585

Return to Risk Ratio

US Treasury Bills Us Corporate Bonds UK Government Bonds Modern US Government Bonds Post-War 19th Century Contemporary Old Masters Global Stock Real Estate Commodities

Global Art Index

-1 0 1 2 A v e ra g e R e tu rn 0 .5 1 1.5 2 2.5 Volatility

(24)

David Kok - University of Amsterdam 24 Then, Table 5 confirms the relations stated by the correlation matrix. There is a negative covariance across art and US and UK government bonds and US treasury. Also, the levels of the variances are in line with the correlation and standard deviation outcomes of the previous results. Further, this is used to construct the optimal portfolio allocations.

Table 5

Variance and Co-variance Matrix

Global Art Index

Old Masters

19th

Century Modern Post-War Contemp orary Global Stock Commod ities US Governm ent Bonds UK Governm ent Bonds Real Estate US Treasury Bills Us Corporat e Bonds Global Art Index 114.195 11.5959 18.181 16.1636 20.71 22.9694 11.8111 29.2493 -13.5657 -12.4422 14.0027 -6.64712 0.196686 Old Masters 11.5959 48.4451 7.25966 7.26697 4.45949 -3.53052 16.8508 25.1894 -5.0628 -2.90468 13.8439 -2.50243 -0.99662 19th Century 18.181 7.25966 26.1325 12.1303 12.1682 15.2591 13.3617 24.3797 -4.22897 -4.6733 10.3278 -1.7205 -1.53407 Modern 16.1636 7.26697 12.1303 13.879 12.381 14.7221 14.437 18.1456 -5.85246 -5.31834 15.6831 -2.90882 -0.34856 Post-War 20.71 4.45949 12.1682 12.381 23.8273 21.9022 11.6693 18.945 -4.10834 -4.37066 14.3592 -1.94578 -1.2576 Contemp orary 22.9694 -3.53052 15.2591 14.7221 21.9022 46.025 11.3402 26.5669 -4.24175 -4.15572 7.01152 -1.89634 1.70423 Global Stock 11.8111 16.8508 13.3617 14.437 11.6693 11.3402 75.056 34.1602 -22.2301 -13.8697 71.876 -12.6786 -2.12873 Commod ities 29.2493 25.1894 24.3797 18.1456 18.945 26.5669 34.1602 164.187 -20.8254 -21.9409 33.028 -10.7761 -1.40083 US Gov. Bonds -13.5657 -5.0628 -4.22897 -5.85246 -4.10834 -4.24175 -22.2301 -20.8254 17.4686 12.3249 -16.8023 9.80367 -0.53965 UK Gov. Bonds -12.4422 -2.90468 -4.6733 -5.31834 -4.37066 -4.15572 -13.8697 -21.9409 12.3249 11.6335 -10.8023 6.83896 0.020396 Real Estate 14.0027 13.8439 10.3278 15.6831 14.3592 7.01152 71.876 33.028 -16.8023 -10.8023 108.403 -9.31377 -1.97334 US Trsy -6.64712 -2.50243 -1.7205 -2.90882 -1.94578 -1.89634 -12.6786 -10.7761 9.80367 6.83896 -9.31377 5.804 -0.1921 US Corp. Bonds 0.196686 -0.99662 -1.53407 -0.34856 -1.2576 1.70423 -2.12873 -1.40083 -0.53965 0.020396 -1.97334 -0.1921 6.59426

(25)

David Kok - University of Amsterdam 25 The optimal portfolio allocation excluding art (Table 6) shows an average return of 1.31, which is not a substantially high figure compared to the individual assets returns. On the other hand, only US government bonds have a higher Sharpe index. This, together with the standard deviation that is lower than any other individual asset portfolio, gives us the financial market benchmark portfolio results. Here, US and UK bonds and real estate is left out.

Table 6

Optimal Portfolio Allocation excluding Art

Portfolio Return 1.309588 Variance 1.536629 Standard Deviation 1.239608 Sharpe 0.124867 Weight Global Stock 0.11326 Commodities 0.025658 US Gov. Bonds 0 UK Gov. Bonds 0 Real Estate 0 US Trsy 0.569455 US Corp. Bonds 0.291627 Total 1

(26)

David Kok - University of Amsterdam 26 Then, in Table 7, shows that in the optimal diversified portfolio, art is included with a weight of 2.87%. Again, US and UK bonds and real estate are left out. In this portfolio, commodities represent the smallest part. This portfolio yields a lower return than the benchmark portfolio and the Sharpe index is lower than the benchmark, but is higher than any other individual asset portfolio. The diversified portfolio outperforms each individual asset class with a Sharpe index of 10.92. On the other hand, the diversified portfolio underperforms compared to the benchmark portfolio that has a Sharpe index of 12.49.

Table 7

Optimal Portfolio Allocation including the Global Art Index

When looking at Table 8, we see that dividing the global art market into various art schools yields a substantially lower Sharpe index of 2.32. Only, the standard deviation is diversified more than when general art is involved. For this reason, involving these subcategories works to diversify the risk. 19th century and contemporary art works are left out. These art schools correlate more than other schools with US treasury bills. The optimal portfolio contains for 78.56% of US treasury and US corporate bonds.

Portfolio Return 1.286109 Variance 1.445599 Standard Deviation 1.202330 Sharpe 0.109210 Weight

Global Art Index 0.028650

Global Stock 0.108946 Commodities 0.020598 US Gov. Bonds 0 UK Gov. Bonds 0 Real Estate 0 US Trsy 0.567027 US Corp. Bonds 0.274778 Total 1

(27)

Optimal Portfolio Allocation including the Art Schools

Because of the overruling part of US treasury involved in the optimal diversified portfolios, there is also looked at optimal portfolio allocation excluding art as an asset class. The benchmark portfolio remains as it is. Art in general is involved in the portfolio with a weight of 2.47% (Table 9). It represents the smallest asset class of the portfolio. Thereby, this diversification yields the highest return of all optimal portfolios. More importantly, the Sharpe index is at 21.82, which almost doubles the benchmark portfolio. On the other hand, this portfolio shows a higher volatility than the benchmark portfolio. Table 10 shows that dividing the art market into art schools, the return of the diversified portfolio is much lower, as is the Sharpe index at 5.74. In this case, only contemporary art is left out due to its positive correlation with other financial assets.

Portfolio Return 1.18224 Variance 1.40380 Standard Deviation 1.18482 Sharpe 0.02315 Weight Old Masters 0.009067 19th Century 0 Modern 0.094010 Post-War 0.011844 Contemporary 0 Global Stock 0.085568 Commodities 0.013906 US Gov. Bonds 0 UK Gov. Bonds 0 Real Estate 0 US Trsy 0.518444 US Corp. Bonds 0.267160 Total 1

(28)

Optimal Portfolio Allocation including the Global Art Index (excl. Treasury Bills)

Table 10

Optimal Portfolio Allocation including the Art Schools (excl. Treasury Bills)

Portfolio Return 1.481442 Variance 2.241451 Standard Deviation 1.497148 Sharpe 0.218175 Weight

Global Art Index 0.024729

Global Stock 0.119984 Commodities 0.044147 US Gov. Bonds 0.204030 UK Gov. Bonds 0.228566 Real Estate 0 US Corp. Bonds 0.403273 Total 1

Optimal Portfolio Allocation

Portfolio Return 1.233675 Variance 1.888536 Standard Deviation 1.374240 Sharpe 0.057394 Weight Old Masters 0.010096 19th Century 0.015260 Modern 0.145643 Post-War 0.023947 Contemporary 0 Global Stock 0.068778 Commodities 0.025904 US Gov. Bonds 0.127663 UK Gov. Bonds 0.241571 Real Estate 0 US Corp. Bonds 0.341138 Total 1

(29)

David Kok - University of Amsterdam 29 On the basis of the optimal diversified portfolio including art, the efficient frontier (Figure 4) confirms it is optimal at a rate of return of 1.28 and a standard deviation of 1.20. This precisely is the point most left on the frontier.

Figure 4

Portfolio Efficient Frontier of Portfolio Allocation including Global Art Index

1.1 1.3 1.5 1.1 1.2 1.3 1.4 1.5 1.6 P o rtfo lio Ret u rn

Portfolio Standard Deviation Portfolio Efficient Frontier

(30)

6. Discussion

In this chapter, the limitations regarding this study will be discussed. With the goal that further research on this topic will lead to more significant results.

First of all, this research involved US treasury into the diversified portfolio. The rate of the US treasury is also used as a proxy for the risk-free rate. This asset class reduces the overall return of the portfolio (Bodie et al., 2011).

Secondly, the Sharpe ratio may not lead to significant results when investments do not have a normal distribution of returns (Bodie et al., 2011). Therefore, other performance measurements can be used to draw more significant conclusions for this study.

Thirdly, transaction costs of art are not involved in this study. There is no data available that contains these costs. As Campbell (2008) noticed, these can become as high as 30% of the auction price. Data containing these costs would give a better representation of the art market figures. Due to the high entry level to the art market, these findings are mostly interesting for HNWI’s or institutions (Campbell, 2008).

Then, the results are based on return indices that were obtained via different methods. The geometric mean is used for the art data and the arithmetic change of returns is used for computing the change of returns over time for the financial assets. This does not necessarily give similar outcomes. It could make the results regarding this long-term dataset, incorrect to interpret (Renneboog and Spaenjers, 2013).

(31)

7. Conclusion

Due to the comparatively bad performance of financial assets in the recent years, investors are searching for alternative investment opportunities. Art as a financial asset has repeatedly been studied with similar results: only the aesthetic-dividend can compensate the investor. On the other hand, art does not relate to traditional asset classes and the art market is booming. This offers diversification possibilities. Art funds assist investors to indirectly invest in the art market.

In this study, is looked at the relation of art with respect to traditional assets and the performance of art as a diversification tool. For this, the global art price index for the period 1998 – 2018 is used as a proxy for the total auction sales in the art market. Also, the index is divided into various schools to check whether some of them show significant outcomes. The benchmark portfolio is created from US and UK financial market data concerning stocks, bonds, commodities and real estate.

The answer to the question whether art can be seen as a profitable alternative for traditional assets is bilateral. First of all, this research shows low correlations between art and US and UK governments bonds, US treasury bills and US corporate bonds. Which points out the diversification benefits of art. To prove this, the optimal portfolio allocation that gives us the highest Sharpe index, at 21.82, included art as a financial asset with a weight of 2.47%. In this case, art is an attractive, but small, addition to the portfolio strategy of an investor. On the other hand, art as an individual financial asset is outperformed by all other financial assets. Art is the riskiest investment and yields the lowest average return. With a return-risk ratio of 0.22, it can be concluded that the return on art investment should purely depend on the aesthetic-dividend of the investor, instead of the financial gain, to compensate for the investment.

Given the limitations of this study, advise for coming research on this topic would be to look for a more complete art market dataset that contains transaction of both dealers, auctions as well as private collectors. Also, a recommendation for the use of more performance measurements has to be made. Investing in art as an alternative investment becomes more and more popular, but still research does not point out its financial benefits. Art funds play a crucial role. More data and information about art funds can make a study on this topic more convenient. Altogether, literature supports the findings of this paper. Directly investing in art should be avoided, when investors are only interested in financial gain. Indirectly investing in art can be profitable, due to the diversification benefits of art as an asset class.

(32)

8. References

Anderson, R. C. (1974). Paintings as an investment. Economic Inquiry, 12(1), 13-26.

Artprice. (2018). 2017 Report on the Art Market. Retrieved on May 21, 2018, from Artprice Website: https://www.artprice.com/artprice-reports/the-art-market-in-2017

Baumol, W. J. (1986). Unnatural value: or art investment as floating crap game. The American Economic Review, 76(2), 10-14.

Bodie, Z., Kane, A., & Marcus, A. J. (2011). Investments. Boston, Mass: McGraw-Hill Europe

Campbell, R. A. (2008). Art as a financial investment. The Journal of Alternative Investments, 10(4), 64.

Chanel, O., Gérard-Varet, L. A., & Ginsburg, V. (1992). The relevance of hedonic price indices: the case of paintings. Universites d'Aix-Marseille II et III.

Gerlis, M. (2017, November 17). Manic Mundi: Leonardo painting smashes records. Retrieved on May 19, 2018, from Financial Times Website: https://www.ft.com/content/8e2b76b4-ca26-11e7-8536-d321d0d897a3

Frey, B. S., & Eichenberger, R. (1995). On the return of art investment return analyses. Journal of Cultural Economics, 19(3), 207-220.

Frey, B. S., & Pommerehne, W. W. (1989). Art investment: an empirical inquiry. Southern Economic Journal, 396-409.

Goetzmann, W. N. (1993). Accounting for taste: Art and the financial markets over three centuries. The American Economic Review, 83(5), 1370-1376.

(33)

David Kok - University of Amsterdam 33 Knight Frank Research. (2018). The Wealth Report: the global perspective on prime property and investment. Retrieved on May 30, 2018, from Knight Frank Website: http://www.knightfrank.com/wealthreport/2018/download

Kraeussl, R., & Lee, J. (2010). Art as an investment: The top 500 artists. VU University Amsterdam, 4.

Markowitz, H. (1952). Portfolio selection. The journal of finance, 7(1), 77-91.

McAndrew, C. (2018). The Art Market 2018. Retrieved on May 21, 2018, from Art Basel Website: https://www.artbasel.com/about/initiatives/the-art-market

Mei, J., & Moses, M. (2002). Art as an investment and the underperformance of masterpieces. American Economic Review, 92(5), 1656-1668.

Pesando, J. E. (1993). Art as an investment: The market for modern prints. The American Economic Review, 1075-1089.

Renneboog, L., & Spaenjers, C. (2013). Buying beauty: On prices and returns in the art market. Management Science, 59(1), 36-53.

Renneboog, L., & van Houtte, T. (2002). The monetary appreciation of paintings: From realism to Magritte. Cambridge Journal of Economics, 26(3), 331-358.

Stein, J. P. (1977). The monetary appreciation of paintings. Journal of political Economy, 85(5), 1021-1035.

Worthington, A. C., & Higgs, H. (2004). Art as an investment: Risk, return and portfolio diversification in major painting markets. Accounting & Finance, 44(2), 257-271.

(34)

9. Appendix

Artprice Dataset, 1998-2018

Artprice Indixes - Quarterly data - Base 100 in January 1998

Copyright © Artprice.com. Glo b a l I n d e x (U S D) Glo b a l I n d e x (EUR) P a int in g s P ri n ts S cu lpt u re s P h o to g rap h s Dr a wi n g s Old Ma ste rs 1 9 th Ce n tu ry M o d e rn P o st -W a r Con te m p o rary USA (U S D) UK (GBP ) Fr a n ce (EUR) 01/01/1998 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 01/04/1998 108 112 104 101 103 91 112 124 106 103 96 103 104 114 102 01/07/1998 107 109 105 106 100 93 114 125 109 103 99 114 109 127 104 01/10/1998 104 105 104 110 97 88 113 106 109 103 98 128 105 120 105 01/01/1999 107 102 106 109 102 85 108 104 112 105 99 123 102 108 107 01/04/1999 103 104 108 109 108 90 101 111 112 107 105 109 108 109 109 01/07/1999 106 113 109 108 111 93 106 100 108 108 109 109 120 122 109 01/10/1999 111 120 111 109 112 98 116 91 108 112 112 113 129 137 114 01/01/2000 117 127 114 109 111 118 122 100 115 112 112 110 127 139 120 01/04/2000 114 131 112 108 110 134 120 107 120 108 111 109 128 136 121 01/07/2000 110 132 110 104 108 122 117 102 118 106 108 114 127 140 123 01/10/2000 111 137 107 99 102 113 117 96 114 104 101 109 124 142 126 01/01/2001 102 134 105 99 103 124 117 90 108 103 104 108 124 145 132 01/04/2001 115 140 106 101 109 132 117 89 106 102 113 115 126 153 136 01/07/2001 103 133 104 99 107 119 114 85 108 101 106 108 124 155 134 01/10/2001 104 131 103 97 102 99 112 81 108 100 102 101 119 151 132 01/01/2002 104 130 104 97 106 98 116 86 103 102 111 99 116 144 132 01/04/2002 112 143 107 100 115 107 118 94 100 106 117 104 120 141 135 01/07/2002 109 132 110 102 111 116 117 91 105 106 122 109 125 144 131 01/10/2002 112 128 112 104 106 127 123 93 111 106 128 107 129 145 127 01/01/2003 120 135 119 109 110 139 135 103 119 112 136 110 138 144 132 01/04/2003 135 140 128 115 118 149 145 106 126 121 150 122 138 147 130 01/07/2003 135 132 131 120 123 142 156 112 122 124 155 130 136 154 120 01/10/2003 138 136 132 122 123 130 161 119 118 125 152 130 139 158 116 01/01/2004 137 129 138 128 129 132 162 122 122 129 166 137 140 158 122 01/04/2004 155 140 145 136 135 141 168 121 132 132 182 161 143 157 128 01/07/2004 148 137 148 138 132 161 164 114 137 134 179 179 145 155 123 01/10/2004 149 136 152 140 129 179 156 117 140 139 179 177 142 156 119 01/01/2005 159 137 160 146 138 179 162 131 147 146 193 181 148 157 124 01/04/2005 169 145 165 151 150 173 173 137 148 149 210 192 160 157 128 01/07/2005 165 147 163 149 148 167 173 137 139 147 215 194 168 159 129 01/10/2005 158 145 159 142 140 161 171 136 131 143 210 190 167 163 130

Art as a financial asset : a performance analysis of art as an alternative investment