chapter  4
27 Pages

## Medieval foreign exchange: a time series analysis

ByADRIAN R . BELL , CHRIS BROOKS AND TONY K . MOORE

Barcelona-Avignon –0.01 1.96 0.20 1.54 –4.45 4.94 Barcelona-Bruges 0.09 13.02 9.35 119.04 –6.45 46.15 Barcelona-Florence 0.00 2.60 0.54 2.82 –3.92 6.80 Barcelona-Genoa 0.01 3.48 –0.53 2.12 –8.42 5.94 Barcelona-Majorca 0.05 0.78 –0.45 3.52 –3.95 2.84 Barcelona-Montpellier 0.00 2.18 0.33 2.41 –4.57 6.11 Barcelona-Pisa 0.02 3.39 –0.10 3.10 –7.03 8.50 Barcelona-Venice 0.02 1.90 –0.12 0.29 –3.24 3.94 Bruges-Barcelona –0.05 2.28 0.01 1.35 –4.99 4.53 Bruges-Genoa 0.00 2.50 –0.28 1.00 –5.64 4.49 Bruges-London –0.04 1.43 0.44 1.54 –4.08 3.85 Bruges-Paris 0.06 0.81 –0.23 2.42 –3.4 3.41 Bruges-Pisa 0.12 1.47 –0.44 1.20 –3.47 2.50 Bruges-Venice 0.12 9.49 2.68 62.85 23.08 30.3 Venice-Barcelona 0.02 1.99 –0.11 1.50 –5.00 4.35 Venice-Bologna 0.00 0.57 0.02 3.63 –2.97 3.51 Venice-Bruges 0.04 2.22 –0.23 2.22 –6.77 4.83 Venice-Florence 0.01 1.07 –0.08 2.78 –3.96 5.19 Venice-Genoa 0.07 1.48 –0.08 1.34 –3.96 3.98 Venice-London 0.14 1.76 0.48 1.50 –2.67 4.44 Venice-Lucca 0.01 0.94 0.10 4.98 –4.32 3.93 Venice-Milan 0.10 5.97 –5.27 48.98 21.85 7.42 Venice-Paris –0.03 1.86 –0.13 0.60 –4.08 4.34 Venice-Pisa 0.02 1.10 0.29 4.09 –3.88 5.40 Venice-Rome 0.20 2.26 1.03 3.81 –3.04 7.03

y, == alDl , +a2D2, +a3D3, +a4 D4, +a5D5, +a6D6, +a7D7, + asD8, + ~D9, + alODl 0, allDll, + a12 D12, + u,

the Aragonese realm. The latter are almost certainly caused by outliers in the percentage change series for these currency pairs as a result of currency revaluations in Bruges (Munro, 2012). In terms of their symmetry or otherwise, some of the series are left-skewed and some are right-skewed; all series are leptokurtic, implying that they have more in the mean, fatter tails but less in the shoulders of the distribution compared with a normally distributed series having the same mean and variance. Again, this feature is very similar to that in standard contemporary asset return series. This leptokurtosis is a key feature of financial time series from a risk management perspective since it implies that extreme movements (i.e. very large swings) are more likely than would be the case under a normal distribution. Barcelona-Bruges is the most skewed and has the highest kurtosis, in particular due to an almost 50 per cent rise between October and November 1390. This reflects a curious delayed impact of the enhancement of the Flemish currency in 1389. The Italian merchants at Barcelona continued to quote the exchange rate with Bruges in terms of the old money until November 1390, when they switched to using the new money (de Roover, 1968, pp. 39-40).