11.11.2014 Views

"Life Cycle" Hypothesis of Saving: Aggregate ... - Arabictrader.com

"Life Cycle" Hypothesis of Saving: Aggregate ... - Arabictrader.com

"Life Cycle" Hypothesis of Saving: Aggregate ... - Arabictrader.com

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Utility Analysis and the Consumption Function 27<br />

seriously biased measure <strong>of</strong> the accustomed level. In the previous test the extent<br />

<strong>of</strong> this bias was measured by the correlation r yy-1 ; the higher the correlation, the<br />

smaller the bias and, therefore, the higher the elasticity <strong>of</strong> consumption with<br />

respect to in<strong>com</strong>e. Margaret Reid and a few others have suggested and tested<br />

alternative methods <strong>of</strong> getting a more reliable index <strong>of</strong> the accustomed level than<br />

current in<strong>com</strong>e and, invariably, it is found that when consumption is related to<br />

such a measure, the elasticity <strong>of</strong> consumption with respect to it rises markedly<br />

above the consumption-current in<strong>com</strong>e elasticity, and <strong>com</strong>es close, frequently<br />

remarkably close, to unity. 46<br />

Another set <strong>of</strong> results that supports our model is that reported in the classical<br />

contribution <strong>of</strong> Dorothy S. Brady and Rose D. Friedman, “<strong>Saving</strong>s and the<br />

In<strong>com</strong>e Distribution.” 47 As is well known, their major finding is that the saving<br />

ratio appears to bear a much more stable relation to the position <strong>of</strong> the in<strong>com</strong>e<br />

recipient in the in<strong>com</strong>e distribution than to the absolute level <strong>of</strong> in<strong>com</strong>e itself. In<br />

other words, the proportion <strong>of</strong> in<strong>com</strong>e saved in a given decile varies much less<br />

over time and space than the proportion <strong>of</strong> in<strong>com</strong>e saved at a given level <strong>of</strong><br />

in<strong>com</strong>e. It is not difficult to see that these results are what one would expect if<br />

our model is correct. As should be clear from the reasoning we have followed in<br />

developing our figure 1.2, the relative frequency <strong>of</strong> households in a given in<strong>com</strong>e<br />

bracket whose in<strong>com</strong>e is below or above their accustomed level, depends, not on<br />

the absolute level <strong>of</strong> in<strong>com</strong>e, but, rather, on the position <strong>of</strong> the in<strong>com</strong>e bracket<br />

relative to the average in<strong>com</strong>e. For example, in a given in<strong>com</strong>e bracket, say<br />

$10,000, we should expect to find a large proportion <strong>of</strong> people whose accustomed<br />

level is less than $10,000 if, say, the average in<strong>com</strong>e is $2,000 and the level<br />

$10,000 is in the top decile; while, in this same bracket, we should expect to find<br />

a small proportion <strong>of</strong> people whose accustomed level is below $10,000 if the<br />

<strong>com</strong>munity average in<strong>com</strong>e were, say, $50,000 so that the $10,000 bracket is in<br />

the lowest in<strong>com</strong>e decile. More generally, it can be shown that provided, as seems<br />

likely, there is a fairly stable relation between average in<strong>com</strong>e in a given decile<br />

and the over-all average in<strong>com</strong>e, then the saving ratio, in a given decile, would<br />

depend primarily on the (relative) short-term variability <strong>of</strong> in<strong>com</strong>e. 48 Thus, if we<br />

<strong>com</strong>pare, over time or space, groups for which the (relative) variability <strong>of</strong> in<strong>com</strong>e<br />

is not very different, the proportion <strong>of</strong> in<strong>com</strong>e saved in any given decile should<br />

be roughly the same for every group. As an example, we can <strong>com</strong>pare the behavior<br />

<strong>of</strong> nonfarm families in different regions and at different points in time (see<br />

Brady and Friedman, charts 1 through 4) and we find our inference confirmed.<br />

Furthermore, within a group, the greater the variability in in<strong>com</strong>e, the greater<br />

should be the variation in the saving ratio as between the lower and the upper<br />

deciles. This inference, too, is supported by <strong>com</strong>parison <strong>of</strong> nonfarm and farm<br />

groups (<strong>com</strong>pare chart 2 with chart 5).

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!