in the great agrarian empires such as Rome, there was a large volume of trade in ordinary goods. Grain, wine, olive oil, ceramics, were traded on a tremenous scale. They were not plundered and then redistributed as Kling supposed - rather, they were in many cases produced for the market.
Second, according to Peter Temin's recent article the Roman empire had a functioning labour market. Robert Allen has done a rough calculation of real wages in the Roman empire in 301 based upon the price edits of the Emperor Diocletian here which indicates that Roman living standards were comparable to those Southern and Eastern Europe in the eighteenth century. Since the Rome empire in 301 had just undergone the crisis of the third century this estimate puts a lower bound on Roman living standards in the early empire.
Temin also provides convincing evidence that merchants had access to sophisticated financial instruments in another article. In plunder based economies there is no need for financial instruments of any kind since goods are acquired through coercion and labour is typically bonded or enslaved.
I disagree with the last point. You need all sorts of sophisticated accounting and financial instruments in order to fund an army or to operate a regime of stationary bandits. Based on my reading of Peter Leeson on pirates I would say that sophisticated political and financial contracts can very well arise in a plunder-based society.
My instinct is to be skeptical of the estimates of large cities in Classical times. One of Gregory Clark's more interesting tidbits in A Farewell to Alms is how innumerate people were.
Think of the necessary conditions for a large volume of voluntary trade in non-luxury goods. On the one hand, you need large differences in productivity in different products in different places, in order for comparative advantage to be strong. In an age where most economic activity is agricultural, this means you need differences in climate and soil. On the other hand, you need transportation costs to be low, which means that trading partners are not going to be hundreds of miles apart. So you need large differences in climate and soil in locations that are close by.
Koyama speaks of "Rome - a city of about 1 million in 50 AD entirely dependent on grain imported from Egypt..."
Why would Egypt have better grain-growing capacity than central Italy? What did Romans have that Egyptians valued in return for grain?
I am open to argument, but I would like to know what the smoking-gun evidence is for (a) the view that Rome really achieved a population of one million and (b) that Rome obtained its grain from Egypt as part of voluntary trade.