Interest rates and national savings in Iceland

Pétur Blöndal, MP of the Independence Party in Iceland, recently claimed that the reason for the high rate of interest in Iceland is that people don't save. In other words, if the people would save more and the amount of savings rise the rate of interest would come down. I beg to differ.

First, let's test Blöndal's hypothesis in the historical context: increased national savings would lead to a lower rate of interest. This implies a negative correlation between the rate of interest and the level of national savings: savings go up, interest rates go down.

The reality, however, glancing at the data (graphed below), is that the correlation between the two different factors is positive, and quite high as well: 0.77 to be specific.

National savings and the rate of interest in Iceland show a positive correlation, not a negative one

Besides the point that this refutes Blöndal's claim immediately, this has to be explained. Why is there a positive correlation between the rate of interest and the level of national savings?

Now, there are three potential reasons for this.

First, the causation is the other way around: it's not savings that influence the rate of interest but the rate of interest pulling savings with it. After all, the higher rate of interest you get for saving a part of your income, the more eager you should be to save. This indicates that changes in the real long term interest rates takes place first and the national savings are influenced, moving thereafter.

This can be tested with a Granger-causality test using the data above, the results of which are the following: the series fail the unit-root tests but they are nonetheless not cointegrated. Not surprisingly then, we see no Granger-causality in the data. Ergo: higher rate of interest does not cause, or lead to, more national savings, according to this dataset. I am therefore going to reject this explanation.

Another potential explanation is Keynes's Liquidity Preference Theory of Interest. Keynes's insight was simple: savings is a two-decision process. First, your decide how much you want to save. Then, you decide how you want to save. If your liquidity preference is high you prefer to save in a very liquid form of savings, such as money or money-equivalents. This you do if you are fearful about the future. If you are not fearful about the future chances are you first of all don't want to save that much - what the heck for anyway, you'll be fine! - and second, you prefer to buy illiquid assets that are not easily used or transferable into money. This leads to a lower rate of interest since the demand for illiquid financial assets goes up and their price as well.

Keynes's LPT could explain both phenomena shown on the graph: if you are fearful about the future you a) prefer to save more and b) you demand a higher price (rate of interest) for departing with the ability of using your existing savings as money to pay for upcoming expenses, perfectly foreseen or not. This would explain why there is a positive correlation between the rate of interest and the level of national savings.

(It complicates things that the rate of interest used in the graph is the 10 year government bond, corrected by the changes in the CPI. This is the "risk free" rate and may well not be the appropriate rate of interest to look at. There is a plethora of interest rates in the economy and many of them are based on the "risk free" rate with a certain, ever-changing, premium set by the market. Looking at the premium might be more informative than looking at the risk free rate but data for this is not easily available in the case of Iceland).

Then, there is the third potential answer, offered by Jónsson (1999) at the Central Bank of Iceland, who, having exclaimed an apparent contradiction in the fact that national savings went down as the rate of interest went up in the 1980s, the following answer: the stock of existing wealth (accumulated saving) offers such high cash-flows in the form of interest that new savings is not wanted. In other words, once the stock of wealth has reached a certain level the rate of interest stops having any significant effect on the level of new savings (which is a flow variable).

This is hard to accept at face value but there is perhaps a grain of truth in this hypothesis. It basically says that people become complacent with what they have once they have reached a certain level of financial wealth and interest rate changes lose their influence on savings. Although it does not explain very well the high correlation between the two time series seen in the graph above, something that Keynes's LPT does, we should remember that Masson, Bayoumi and Samiey (1998) found that increased income (GDP per capita) lead, in low-income countries, to an increase in savings while, in the case of high-income countries, it lead to a decrease in savings. So, effectively, that people become happy with what they have once they've reached a certain level of savings is not an unlikely explanation. But, again, Jónsson's idea hardly explains the high and positive correlation between the rate of interest and national savings.

Obviously, looking at this matter without considering e.g. the level of income is incomplete. Lower level of income may force national savings to go down since people don't have anything left, after their everyday expenditures, to save. But whatever the reason is - curious minds with more time on their hands than I are encouraged to try and find out - the lesson is that Blöndal is certainly not right: increasing national savings would not lower the rate of interest, except, of course, completely coincidentally.