F. Velde ha
publicado "On the Evolution of Specie: Circulation and Weight Loss in 18thand 19th Century Coinage", y en su resumen escribe: "I measure the
parameters of coin wear using data collected in the 19th century. A comparison
across denominations and countries shows that coin wear (in relative terms) is
linear in the logarithm of coin value. Data from coin hoards of the 18th and
early 19th centuries yield similar estimates of mean coin wear, showing that
hoards provide useful information. Finally, under assumptions of normality for
initial coin weights and coin loss I use maximum likelihood estimation to
recover the parameters of the wear process from a sample of coins whose age is
unknown. The method performs well on the hoard data (for which the age is known
and can serve as a check)". En sus conclusiones leemos: "My
results confirm, extend, and modify those found in the existing literature.
Coin wear (in commodity money regimes) can be modeled as a Brownian motion,
with weight loss per unit of time approximately normal with constant mean and
variance. The difference in coin material between silver and gold coinage is
not quantitatively important in explaining variation in coin wear across
denominations. The appropriate measure appears to be relative weight loss, not
absolute weight loss. Under that transformation, a striking relation appears
between annual coin wear and the log of value. If coin wear is taken as a proxy
for velocity, then velocity is inversely related to value. That seems to be the
right transformation: when absolute weight loss was regressed on value, the coefficient
was positive, with the counter-intuitive implication that higher denominations
are more intensively used. This relation appears stable across 19th century
countries. This suggests that considering only quantities of denominations
produced is not a good way to assess whether small coins were produced in sufficient
quantities for the needs of trade, since velocity varies significantly across
denominations. Coin hoards are a good proxy for the general circulating medium,
at least with large samples (a thousand or more). Coins from the numismatic
trade offer good data for estimating coin wear, but the selection problem can
be an attenuating factor. Further work could be done to investigate the extent
of this bias. For older (and scarcer) coins, that bias is likely to be less
important. Coin wear also appears to incorporate a spatial dimension. Finally,
I have derived an estimator of the parameters of coin circulation processes
(mean and variance of weight loss, and rate of coin loss) that does not rely on
knowing the vintages of the coins. This method could have wide applicability to
hoard data from ancient and medieval periods.