Friday, February 16, 2007

The Spectre of Inflation - Part II

As pointed out by George in yesterday's comments, the CPI might not apply as well to your personal situation if you don't buy the entire basket of goods that it covers. For example the CPI has mortgage interest rates included, but if you own your home this doesn't apply to you. So this does tend to mess up the results a bit if your looking at your personal rate of inflation.

As I mentioned in closing of yesterday's post, the Bank of Canada admits the CPI bias tends to be about 0.5% high to real inflation. What is interesting is the IMF mentions that our inflation bias tends to run 0.5 to 1%, which is still better than the US which runs over 1%. So if you look at the data from 1995 to present the core inflation rate averaged a mere 2.0%, which a maximum of 4.6% in Feb 2003. So if you assume a 0.5% bias is correct we actually only averaged 1.5% inflation during the last 12 years.

During the my research for these posts I came across an interesting paper that looked at seniors spending in relation to the CPI. It analyzed data from the 1970's and 1980's and found the that the CPI overstated inflation by 50% for seniors. Which means that 12.1% inflation back in 1981 was actually closer to 6% for seniors during that time. The paper does go on to state the CPI was fairly accurate during the 1990's, but the 0.5% bias still applied.

What is perhaps the most interesting about inflation is what the Big Five Banks use for their default inflation number is all their online retirement calculators.

TD = 3%
RBC=2 to 4%
Scotia = 3%
CIBC = 3%
BMO = 2.7%

Doesn't that seem odd that if the CPI has averaged 2% for the last 12 years with a 0.5% high bias that all the banks are near 3%? I thought so. So what does that mean if you cut your inflation estimate in half to 1.5%? Then you would need a lot less to retire. Yet a word of caution here, if you cut back your estimate here, you should increase your spending estimate with some extra padding. You need to have some extra in your calculations to account for various things that can go wrong, but I often find most people get a little crazy about padding the numbers. So much that they end up with such an over padded estimate that you might as well be living in a bubble than facing real life.

Of course a conclusion like that will have disbelievers, saying 'what if inflation goes up to 9%?' Well I would expect my yields to follow suit. After all if your truly frightened of spectre of inflation you can always invest in Real Return Bonds from the government of Canada. They will adjust your base amount according the CPI and still give you a couple of points interest beyond it. I would caution anyone who is interested to do your homework first on these bonds are they are taxed in a rather unique way.

Have a good weekend,

5 comments: said...

Hey Mr. Dream,

Great post! Personally, I like to err on the side of caution which is why I personally like to use higher inflation numbers like 3%.


Canadian Dream said...


I totally respect that everyone will calculate their number in a different way and use different assumptions.

I just like to point out issues with those assumptions and let people make up their own mind.


Ro said...


Canada's inflation has averaged 2% since the Bank of Canada made inflation targeting the aim of its monetary policy, 15 years ago.

Although inflation targeting has worked extremely well over this time frame, there is still the danger that in the future Canada might have to change its monetary policy (we have done so numerous times in the past).

Given that the time frame on a retirement portfolio is 20-30 years, it might be wise to stress test with different inflation scenarios.


Investoid said...

Another factor which may lead to overstating inflation in retirement is the 14.79% weighting that StatsCan has on non-ownership shelter costs (ie. rent, mortgage interest and replacement cost).

If you're a retiree who's paid off their home and plans on living there, then these costs are non-existent. As the most recent StatsCan release shows, these items can be the major factor in CPI increases. When you take those out, you have even less to worry about.

Canadian Dream said...


Stress testing with different numbers may not be a bad idea, after all if I can retire at 45 I'm looking at a 40 year retirement span.


A great point on the CPI. I didn't get into any details of a the actual makeup of the CPI since I was already at two fairly long posts.