> I may be missing something obvious here,
> but isn't there a third factor that can
> make the profit rate go up - a decrease
> in the capital-output ratio? This wouldn't
> require the stock of capital to shrink,
> it could be achieved simply through
> technological improvements that increased
> the amount of output yielded by each
> additional increment of capital (holding
> labor input constant). In fact, the
> capital-output ratio declined a lot in the
> 90's, which is why the profit rate was able
> to rise without a significant change in the
> profit share of national income ("the rate
> of exploitation") or a shrinking capital
> stock.
There are n ways to skin this cat. The decomposition of r that you suggest is this:
r = s/K = (s/Y)*(Y/K)
where r is the profit rate, s is the flow of surplus value, K is the stock of capital, Y is the value of output (flow). You can define s/Y as the "profit share" and Y/K as the "average productivity of capital" (the reciprocal of the "capital-output ratio"). Profitability can be restored by expanding the profit share and/or by lowering the capital-output ratio. A problem with this decomposition is that, in a crisis, Y's components can move in different directions and each of those effects will not be apparent in the formula.
This decomposition of r may be clearer in making the point:
r = s/K = (s/v)*(v/K) = e/h
where v is the flow of variable capital (wages), e = s/v is the rate of exploitation, and h = K/v is the capital composition (if the wage rate is assumed constant, h equals the capital-labor ratio). Then the profit rate is a Cobb-Douglas function with increasing returns in e and (1/h). More simply, a positive function of the rate of exploitation and a negative function of the capital composition.
To increase r, you need a higher e and/or a lower h. To lower h you either increase v while holding K constant, shrink K while holding v constant, or some combination of them. Note that v = w L, the average wage rate (per unit of labor power) times the amount of labor power utilized. So h = K/L*w = k/w, where k is the capital-labor ratio. In a crisis, with unemployment, both L and w are hit. That's why the immediate likely scenario is K shrinking, not v going up.
On the other hand, e goes up if s expands as a proportion of v. In a crisis, again, v is likely to shrink because, with unemployment, both L and w tend to drop. However -- as you correctly point out -- another way for the rate of exploitation to increase is by increasing productivity, which has the effect of cheapening the wage basket. Thus, the workers' standard of living and work day (and labor intensity) can stay the same, yet s will go up.
I neglected to mention this way of expanding e, because -- again -- one tends to think of unemployment (and a shrinking workers' standard of living) as the most immediate likely scenario in a crisis. During a typical recession, labor productivity declines.
Yours is a good point to raise, because over time, with the cheapening and concentration of capital (plus the insecurity of workers), it becomes possible for capitalists to reorganize production and boost productivity. Also, with a strong, well-directed, and well-managed push in public spending, average productivity could be increased or, at least, the conditions for its further increase could be laid out. (The latter requires that total K stay fixed or doesn't increase so much as to offset the increase in e.)
So, those would indeed be other ways for the rate of exploitation (and the profit rate) to be restored.