Pastels, Piketty and the picayune

Several years ago I put up a page on my pastellists website reporting prices achieved by pastels at recent sales. I didn’t spend a lot of time analysing the numbers: they are prices, not values, nor did I – until this weekend, when I went back to take another look – investigate what really interested me, namely what had happened to prices over the long term. This was a problem I’d dismissed before as too difficult. If you own a picture which once sold for 1 million francs, you may understand why – the date of that sale was 1957, when the franc wasn’t worth what it used to be.

Anyone who is interested in the art market will be aware of a great deal of research in this area. There are books like the 1961 classic The Economics of Taste by Gerald Reitlinger; experiments by the British Rail Pension Fund, whose investments in fine art for two decades were disposed of in 1995; presentations by art advisors directed at persuading the newly rich to buy; and of course innumerable statistical databases at wildly varying subscription rates. Needless to say none is specifically devoted to pastels as such; and none can overcome the fundamental problems with such analysis.

The most obvious of these is that no two pictures are the same. I can recall from my days as an investment banker trying to establish a property index from which derivatives might be traded: we could never eliminate what we called the basis risk. It is not just that a prime office building in Mayfair doesn’t behave the same way as a Northern industrial warehouse; two very similar buildings of the same type won’t always go for the same price. And in pictures that is true in spades.

The second problem may seem trivial, but it isn’t: currencies. Of course you know that there were lots of different coins in the eighteenth century, some so hopelessly obscure that they provide titles for blog posts. And if you spend a few hours looking into this you’ll see that the multiplicity of coinage and conversion rates was a good deal more complicated than you might think. The real problem was that coins converted at available exchange rates had very different purchasing powers. London, then to a far greater extent than now, was vastly more expensive than anywhere on the Continent. Oddly that distortion is more difficult than fluctuations in rates per se.

But of course the biggest obstacle in this analysis is inflation. If a picture cost £20 in 1750, what is the equivalent today? Is it the number of loaves of bread you can buy for the same amount? Is it the annual pay today of an occupation that earned £20 p.a. in 1750? Or is it the weekly pay today of someone earning £20 a week in 1750? (a very different answer). What these art studies show is that luxury goods in 2015 cost relatively more than they did in 1750: but you have to be careful not to be circular if you try to invent a luxury goods index to inflate your prices. However imperfect, the best you can do is to use the general consumer inflation index produced by the Bank of England and backdated to 1750.

How good is your intuition about those numbers? How much do you think £1 in 1796 is worth in today’s money? What about 1914? It may surprise you to learn that the answers are the same: both are worth about £100. Of course there were fluctuations over the centuries, but not by orders of magnitude – until the first world war. Inflation is essentially a modern phenomenon. And if you think it has been banished, what do you think £1 in 2000 would be worth today? The answer – £1.50 – may surprise you. Next time a politician complains because inflation is below 2%, challenge him. There is nothing good about inflation at any level.

Now combine currency exchange and inflation rates. Take 1 franc in 1901. What was it worth in 2001? You don’t in fact need to worry too much about the black hole into which the franc disappeared in 1960; you simply need to compare the basket of goods you could buy in the two years. The official French inflation data provide the answer of €3.14. If you’d converted that to sterling in 2001, you’d have got £1.95 in 2001 money. But what if you did it the other way round? 1 franc in 1901 was worth about 10d.; adjusting that for British inflation over the century (73 times) provides £2.92, about half as much again.

This illustrates another methodological problem, although the numbers aren’t as far apart as I thought they might be. And indeed for the purposes of the trends I wanted to consider, such inaccuracies are minor. The scale on which the prices of pictures move in response to fluctuations in taste is logarithmic, and only orders of magnitude need be compared.

So here are a few reflections that struck me as I was working with the historical numbers you can find in my revised Prices page together with a commentary which I won’t repeat.

Firstly, prices were much more closely bunched during the eighteenth century than they are today. This is not just an effect of comparing secondary market figures today against origination prices: I suspect that the dispersion of prices of today’s professional artists (in our winner-takes-all world) is far greater than 250 years ago. There were of course lots of incompetent jobbing artists whose advertisements for pastels at a guinea each fill the pages of the Dictionary, but they need not concern us in this analysis. All the serious pastellists whose work is recognised today seem to have commanded prices of the order of £15–25 or so for a head and shoulders portrait. The very best might occasionally have been able to charge say twice that amount: but in today’s money that would barely exceed £10,000. The only exception was the famous 48,000 livres La Tour demanded, but did not receive, for his famous portrait of Mme de Pompadour, a sum worth mid-six figures today.

Second, with pastel the dip in prices throughout almost the whole nineteenth century was extraordinary, and the pickup at the end of the century dramatic. You can see this with the figures on the Prices page – although of course there is already a bias in a data set defined by a price threshold. One of the features of the analysis is the widening gap between the top and the middle, in parallel with Piketty’s findings on overall wealth. But the trend is also confirmed by working through many further mid-ranking examples: to take a single example, a Perronneau which sold for Fr48 in 1867 (equivalent to £200 today) and for Fr70,000 in 1929 (£150,000). Today at auction it might fetch a fraction of that.

One can also see the role of fashion in this. An artist who has been completely forgotten will be “plucked from obscurity” by a single auction record, as instanced by the 1250 guineas paid for a Gardner in 1908, leading to a sudden demand for his works as he joined the superleague of then fashionable names of La Tour, Perronneau and Russell. Four years later the 600,000 francs paid for La Tour’s Duval de l’Épinoy (Museu Calouste Gulbenkian; above), which incurred the wrath of the editor of the Burlington Magazine, set a level which has never been exceeded, at £2.4m in today’s money (and of course far more if inflated in line with luxury goods). Curiously Liotard, who dominates the list of modern prices (taking half the top twenty places), was not part of this club: it was only after the private purchase by an American museum in 1986 that his prices took off, with Geneva paying a sum equivalent to £900,000 today for Lady Tyrell at auction in 1991.

The depressing fact is that no sane investor would choose eighteenth century pastels as a home for their money. You need to be not a cynic, but (to complete Wilde’s text) a sentimentalist – “one who sees an absurd value in everything, and doesn’t know the market price of any single thing.”

Advertisement