So, too, as to the indirect effects of a general fall in the price level. No one description can be given that will hold true of all instances. If the main cause at work is of the kind that may be called "natural," for example, a gradual increase in the productivity of industry, or a decided falling off in gold production, such periods are not necessarily periods of depression in industry. Employment may be constant and weekly and yearly earnings high. Thus the period of 1873-1896 in the United States was one of declining prices and it is generally admitted that that period was one of great industrial activity.[54] Moments of excessive activity are rarer in periods of falling prices than in periods of rising prices, but the average amount of unemployment may be either greater or less. Again, if the decline of prices is in reality a movement from a state of depreciated paper money to a gold standard, there is a possibility that the period may be one of industrial activity due to a prevailing confidence in a coming recovery. It is more likely, however, that such a period will be characterized by a falling off in business activity and an increase in unemployment, particularly at its commencement. Lastly, if the price movement is an indication of such a period of depression as may precede and usually does follow serious industrial crises, it is ordinarily accompanied by liquidation and curtailment of production. In these periods, and especially at their height, unemployment grows and earnings fall more than wage rates. Or wage rates may remain comparatively steady, but weekly and yearly earnings will fall. The extent to which this fall in earnings will go depends upon the seriousness of the industrial maladjustments.[55] Still it is safe to conclude that a period of serious depression following upon a crisis is the least favorable phase of the industrial cycle for the wage earners--notwithstanding the fact that wages frequently fall more slowly than wholesale prices, and somewhat more slowly than retail prices. 5.--Our object in discussing the effect of price movements on distribution is to discover how they complicate the problems of wage settlement. Before proceeding to this main purpose, however, it is desirable to pay particular attention to one doctrine of the relation of wage change to price change which figures prominently in current discussion.