What's all the hoopla? Aren't we just talking about marginal analysis? That's a been a standard part of the economic toolkit for over a century (among others, Henry George described marginalism in 1897, in The Science of Political Economy). It occurred to economists that the overall effect of adding one more unit to the whole pile of something is a rich piece of data, useful for studying how people make complex choices. Here's an example from econlib.org:

A bus that is half-empty can take on more riders with zero or very little extra cost perhaps just a few cents more for wear and tear and the cost of gas to haul an extra 150 pounds. Economists would say the marginal cost of an additional rider is nearly zero. But, if buses are always running packed with lines left standing, then the marginal cost of additional riders would be the entire cost of adding another bus. It is very common to have to compare different marginal costs for different scenarios in order to decide which alternative to pursue.

Applying marginal analysis to workaday business decisions is relatively trivial. But the margin of production applies to the entire society! In this case, the "one more thing" being added is the next piece of land to be placed into production the best land that can be found that isn't yet owned, and therefore doesn't command any rent. Labor and capital get to keep all the wealth they produce on that site. Therefore, they will not agree to work elsewhere for any less than what they could earn working for themselves, where the land is free.

The margin of production is the best available opportunity, in a given place and time, for labor to employ itself without the payment of rent. The level of wages available to labor at the margin of production determines the general level of wages in society.

What do we mean by the "General Level" of wages?

There are, of course, many wage levels in society. Migrant laborers might work long days in the hot sun for a bare-subsistence wage; movie stars might make tens of millions for acting in a film. Between those extremes, there is a huge variety of wage levels.

We can describe overall wage levels in society in various ways. We could take a mean average: simply divide the total amount of wages by the total number of workers. But, if there are a few people who get extremely high wages, this average could seem misleadingly high. A more representative average figure might be the median: the figure at which exactly half of the people make more money, and half make less.

But if we are seeking to understand the cause of persistent poverty, averages are not the most important thing. If the wages of professional and corporate workers were to increase sharply, the average wage level could go up, even while most people's wages fell! It is vital that we keep the purpose of our inquiry in mind: Why is there poverty in the midst of plenty?

The wage figure that most decisively influences the distribution of wealth is that of the lowest-paid worker: someone who is willing and able to work, but who has no special skills, no union membership, no other particular bargaining power. The wage earned by this kind of worker is what we mean by "the General Level of Wages." This general level is determined by the wage that one of these unexceptional workers can earn, rent-free, at the margin of production.

Here's another way to see it: there is a market price for the labor of an unexceptional worker. Market prices, as we know, depend on supply and demand. If workers have a viable option for self-employment at the margin of production, then the supply of workers available to employers is decreased. The cost of hiring basic workers a.k.a. the general rate of wages has to go up. Conversely, if workers have no viable alternative for self-employment, the supply of workers increases, and employers will have to pay them less.

Today's economy is characterized by there being fewer jobs than there are workers eager to fill them. There is a structural surplus of workers, and that's why the general rate of wages tends to be so low. There are many complex ramifications and special cases but the basic principle, the Law of Wages, is that simple.

In its simpler manifestations, this law of wages is recognized by people who do not trouble themselves about political economy, just as the fact that a heavy body would fall to the earth was long recognized by those who never thought of the law of gravitation. It does not require a philosopher to see that if in any country natural opportunities were thrown open which would enable laborers to make for themselves wages higher than the lowest now paid, the general rate of wages would rise; while the most ignorant and stupid of the placer miners of early California knew that as the placers gave out or were monopolized, wages must fall. It requires no finespun theory to explain why wages are so high relatively to production in new countries where land is yet unmonopolized. The cause is on the surface. One man will not work for another for less than his labor will really yield, when he can go upon the next quarter section and take up a farm for himself. It is only as land becomes monopolized and these natural opportunities are shut off from labor, that laborers are obliged to compete with each other for employment, and it becomes possible for the farmer to hire hands to do his work while he maintains himself on the difference between what their labor produces and what he pays them for it. Progress and Poverty