My deepest concern was the irrigation of desert.
I have not ordered, nor was I able if I wanted, construction of a granary in Mr. Roger's. FYI - I lack access to specific info on any of those cities in mention - mainly because I have little sparetime to get fully involved with the game at current, but also because finding updated screenshots of an area has proven somewhat difficult.
And now for my thoughts on efficiency:
if you produce a given amount of output, say x units of a good, and you minimize costs - e.g use an optimal amount of input (lowest required amount of input for the given output), that might be considered effective production.
Second, if you produce x units, and the demand for that good is lower, say x-1, then the amount produced that exceeds the demanded amount could be considered ineffective production - in all the time resources are being spent on something that is not demanded, when those same resources could have been applied to produce other goods.
Thirdly, in civ3 context, I reckon that the Kaldor-Hicks criteria for efficiency could be applied: If terrain is improved in one sector that is relatively high in productivity, for it to become more productive, with the result that some resources (here workers) have to be taken from elsewhere to do so, resulting in a slight loss for the sector those resources where withdrawn from, then the descission was effective if the net benefit of the nation as a whole is larger than the reduction for the specific sector.
You could also apply the Pareto-criteria: Assume you have 2 cities, city A and B. They have the same amount of worked tiles, and the same growth rate. You then improve 1 tile in city A, whilst B has no changes. Net result is increased efficiency - as city A is more productive, whilst city B has not suffered a loss.
Then there is the cost-benefit analysis:
If the same cities A and B faces the descission of whether to improve a tile in A or in B - the planners have to evaluate which improvement that gives the greatest benefit (benefit-cost=net benefit). It follows that the most efficient alternative is the one that gives the largest net benefit
This was today's lesson in microeconomics for civplayers.