Clara's original question asked about the efficiency of slavery vs. free labor, among other things. I've recently finished another book by Gavin Wright that touches on the issue.
Wright argues that economic evidence on the efficiency of American slavery, as a labor system, is inconclusive. One the one hand, there is little evidence demonstrating that slavery was less efficient than free labor. On the other hand, the evidence for economies of scale due to gang labor (the basis for most such claims) is also iffy at best.
Wright argues that slavery's advantages lay elsewhere. American slavery was not simply a system of labor organization. It was also a set of property rights. It was the property rights aspect of slavery that made slavery tremendously attractive and perpetuating.
Before the War, slaves constituted, by far, the single most valuable category of non-real property in the United States. Unlike real property, slave property was transportable. The characteristics of slave property meant there could be a region-wide market (wherever slavery was legal), giving slaveholders ready access to credit.
Looking at settlement patterns in the North and South, what is distinctive is that northern territories developed slowly at first, as immigrants gradually moved in from the East, generally buying small plots that they could afford to purchase from savings. In the South, however, when new territory opened up for settlement, existing slaveholders were generally able to use agents and credit to buy larger pieces of the best land. Even without moving there, existing slaveholder could then arrange for the transfer of existing slaves there, or the purchase of new slaves (more credit), to clear and begin planting. In short, slavery gave owners the ability to quickly take advantage of economic opportunities in distant and even otherwise undesirable locations.
Second, the property aspects of slavery allowed owners to allocate slave labor as they wished (with some exceptions, eg, Sundays off), and not in accordance with the wishes of the slaves. For example, slave women often worked in the fields. White plain folk women may have done so occasionally (see Stephanie McCurry), but certainly the hope and expectation was that they would do domestic chores. In short, "Slaveowners could reallocate family labor from household or nonmarket activities into cash crops, using profitiablity criteria."
Third, in the north, employment at will quickly became the rule, which meant that free labor could leave at any time. The property aspects of slavery meant that slaveowners could be sure that their labor was captive and would be available at peak periods. This meant that slaveowners could plant increased amounts of cash crops. This allowed slaveowners to maximize returns, particularly in exceptional crop years like 1859.
If you figure in the value of wealth represented by slaves themselves, the south was just as wealthy as the north in 1860. The wealth manifested itself in different ways, of course (basically slaves vs. land), and slavery had all sorts of ramifications (on land settlement patterns, tool usage, schools) that amounted to hidden liabilities, but that is another story.
To recap, Wright argues that "productivity" is a red herring. Whether slaves sowed or reaped a field as efficiently as free workers was only a very small part of the reason that slavery was or was not attractive as an economic matter.
Amazon.com: Slavery And American Economic Development (Walter Lynwood Fleming Lectures in Southern History): Books: Gavin Wright