Inflation can be the silent killer of retirement plans. Like high blood pressure, its effects are often unnoticed until it is too late. Individual investors must explicitly allow for inflation into their planning or risk being very disappointed in their retirement.
The problem is that inflation shrinks the purchasing power of a dollar over time. For example, if one had saved a dollar in 1977 and kept it without investing it, that dollar would be worth just 28 cents today. That is to say, its purchasing power would have gone down by 72%. That is because inflation averaged 3.5% per year during this period. Lower rates of inflation are less harmful but higher rates are possible, and could devastate one’s savings in a hurry.
What is inflation?
The Consumer Price Index (CPI)
Inflation is defined as a general increase in the level of prices. That does not mean that every price is increasing, of course. If a few prices are decreasing, while the majority are increasing, it is likely that inflation is present. This raises the question of how we measure inflation.
The most common measure is the consumer price index, or CPI. The CPI looks at a basket of goods and services typical of the average family’s expenditures each month. It compares the total expenditure required this month to purchase this basket, to the total expenditure that was required to purchase the same basket last month. The ratio of the two amounts shows the percentage increase month-over-month of the average price faced by the average family. This monthly increase can be expressed as an annual equivalent, or “annualized”, by multiplying by twelve. Alternatively, and more commonly, an annual inflation number is obtained by comparing the expenditure on the basket of goods and services to the expenditure twelve months ago.
This approach works well when we are measuring inflation over a reasonably short period of time. However, it has to be supplemented if longer periods of time are involved. Family expenditure patterns change over time, and the average family this year may no longer be buying the same things, or in the same proportions, as two years ago, much less ten years ago. To solve this problem, a new basket of goods and services is created each year, based on an annual survey of households. Each basket is used to compare two immediately adjoining years. The increases from each comparison are then compounded over time.
For example, a basket was designed based on expenditure patterns in 2003. This basket is used to compare expenditures, and so inflation, in 2004 relative to 2003. The result was 1.8%. Next, a basket was designed based on expenditure patterns in 2004. It is used to compare expenditures in 2005 relative to 2004. The result was an increase of 2.2%. Inflation from 2003 to 2005 is then obtained as (1 + 0.018) * (1 + 0.022) = 1.0404, which is generally expressed as an inflation rate of 4.0% (rounding). In this way price levels can be compared over time even though expenditure patterns are changing.
Other common measures of inflation
The above measure is called the total or “headline" CPI and is published monthly by Statistics Canada, the Bureau of Labor Statistics in the United States, and central statistical agencies in other countries of the world.
Statistics Canada also publishes a measure called core CPI as defined by the Bank of Canada. It is the same as the total CPI, but omits eight components from the basket, and hence omits price changes in these components from the measured price change. The two main components omitted are energy and food. This measure of inflation is relied upon by the Bank of Canada when it designs its monetary policy, e.g. when it sets interest rates on short term Government of Canada bonds. It is not used for any other official purpose, contrary to the beliefs of some.
The Bank of Canada prefers to monitor the core CPI rather than the total CPI because it finds the total CPI too volatile from month to month. Energy in particular can fluctuate wildly in response to oil prices on international markets. Nevertheless, over longer periods, e.g. several years, the total CPI and the core CPI give roughly comparable results.
The CPI measures inflation from the point of view of households. Other points of view are possible. For example, Statistics Canada publishes a Producer Price Index, which measures price changes for inputs purchased by enterprises, e.g. raw materials, machinery and equipment, and so on. Other indices measure inflation for commodities, or for specific sectors of the economy. Statistics Canada also publishes a broad index of inflation, called the GDP Implicit Price Deflator, or GDP-PI. This measures inflation in the economy as a whole. It is a by-product of producing the national income statistics, and the weightings of various goods and services are the same as in the composition of the GDP.
These alternative measures of inflation are of very limited usefulness to the individual investor, except perhaps as context for the CPI. (An exception is the GDP-PI, used by the CRTC each year to index the price of land-line residential telephone service in those localities where prices are still regulated because there isn't much competition.)
Limitations of the CPI
As stated above, the CPI measures inflation as experienced by the average household. Inflation for an individual household can differ from this average, in a few cases significantly. However, in most cases, the average is close enough, and total CPI is widely used for financial planning purposes. If an individual investor believes that he or she might be a special case, it is relatively straightforward (although sometimes tedious) to identify the expenditure weights that apply in his or her case. These can then be compared to the weights in the Statistics Canada basket, if the weights differ, Statistics Canada sub-indices can be used to construct a CPI specific to that individual.
At various times, there have been claims that the CPI, an average, does not properly represent the experience of certain key sub-populations. In particular, it has been said that the CPI understates inflation as faced by older Canadians, or by the poor, or by non-traditional families. However, historic results suggest that, while the total CPI may misrepresent inflation for any given group over periods of two to four years, over longer periods of time these differences disappear. (Note that the problem of older households is more acute in the U.S., where health care expenses are an important and growing item with age, This problem is much smaller in Canada, where the majority of health care is paid by universal health insurance.)
Treatment of shelter or housing can be controversial. This stems from the fact that housing is both a service (shelter) and an investments (a real property). An increase in the price of houses reflects in part the increase in the price of shelter and in part gains on a real investment. As a result, there are two different approaches to the inclusion of shelter in the CPI.
The first approach involves including rent where a household rents, and an equivalent rent where a household owns the house in which it lives. The second approach involves working out an annual cost of housing based on the prices of houses, interest rates on mortgages, and maintenance and operating costs. Note that this second approach tends to be more volatile than the first, but corresponds more closely to reality for households owning their own houses. Central statistical agencies have switched back and forth between the two approaches over time, making for some confusion. Currently, Canada uses the second approach.
Note that this is part of a more general issue. The CPI is a consumer price index, and is intended to measure changes in prices of what households consume each year. It is not an asset price index; i.e. it does not attempt to measure changes in the prices of assets such as real estate, plant and equipment, precious stones, etc. Some observers believe that an asset price index would be very useful as a guide for monetary policy, but that is a quite different issue and of limited interest to individual investors.
Another controversial issue is the treatment of quality improvements in products. This is particularly acute for information technology products, where improvements typically come in the form of more functionality rather than lower price. Statistics Canada, following the example of the U.S. Bureau of Labor Statistics, adjusts prices to take into quality improvements, to the degree it can.
To take a very simplified example, if a computer with one terabyte of memory sells for $1000 and the same computer with only 300 megabytes of memory sells for $800, the analyst will assign a value of $200 to the extra 500 megabytes. Suppose now that the next year these two computers are no longer offered for sale, but are replaced by a new model with 1.5 terabytes of memory, at a price of $1100. The analyst will value the new model at $1200 for the previous year ($1000 plus $200 for the extra 500 megabytes). Comparing this imputed $1200 last year with the sale price of $1100 this year, the price of a standardized computer has dropped by 8%. This is even though the price paid by the consumer has actually gone up by 10%, from $1000 to $1100; for his $1100, the consumer is getting 17% more computer.
Such adjustments for quality, obtained by inferring what consumers might pay for a better quality product, are called hedonic price adjustments. Some observers object to them, pointing out that the consumer has to spend $1100 in the second year, even if he or she doesn't want to extra 500 megabytes of memory. They also object to a variety of technical changes, with much smaller impacts, that have been introduced over the years. The most vocal of these objectors has been the website Shadow Government Statistics.
Critiques of officially published CPI numbers also allege that governments have political reasons to manipulate reported inflation numbers. Many government benefits, such as Canada Pension Plan (CPP), Old Age Security (OAS), public sector pensions, as well as income tax brackets and deductions, are fully indexed and grow as the CPI grows. The argument is that, to reduce outlays on benefits and to increase tax receipts, governments understate CPI.
The counterargument is that the methodology used by Statistics Canada is reviewed by an advisory of independent experts, including academics and participants from the private sector. The entire process is very well documented and perhaps more transparent than any other government process. As well, senior public servants and Members of parliaments and Cabinet Ministers, once retired, enjoy pensions indexed by the very same total CPI.
The consensus among public and private sector and academic economists in Canada is that, rather than understating inflation, the CPI actually overstates it by some 0.5% per year. This is due to failure to completely adjust for introduction of new products, quality improvements, and increased purchases from low-priced outlets, including online shopping.
Measuring expected inflation
The CPI and the other indices mentioned above are used to measure historic or actual inflation, as experienced by the average household in Canada. This is useful for analyzing past investment results. However, it is of limited use for financial planning.
In financial planning, the important measure is what inflation will be over the planning horizon. Investors saving for retirement or other distant goals need to think in terms of their purchasing power. They will be spending dollars that will have been eroded, to some extent at least, by inflation. (Strictly speaking, a financial plan should be in after-tax and after-inflation dollars, or “net-net”. The after-tax aspect is discussed in another article.)
The financial analysis can proceed in either real terms or nominal terms. A real analysis expresses all revenues and expenses in “constant dollars”, i.e. dollars adjusted for inflation. For convenience the current year is used as a base, i.e. all revenues and expenses are in 2013 dollars. This allows the investor to more easily base his projected expenses on expenses today. These will be increased or decreased to reflect “real” changes in spending patterns, e.g. the need for increased health care, or travel in retirement. On the revenue side, some revenues are already in “real” terms, e.g. CPP, OAS, Guaranteed Income Supplement (GIS) and federal public service pensions are adjusted annually for inflation. Thus the corresponding revenues will be constant over time in real terms.
Other revenue sources, e.g. non-indexed or partially-indexed pensions, investments in stocks and nominal bonds (bonds that pay interest that is not adjusted for inflation), rents from real estate, are in “nominal” terms, i.e. they are in conventional before-inflation dollars. These must be discounted by the expected inflation between now and the date when the money will be needed. Thus we need an expected inflation rate.
More often, the financial plan proceeds in nominal terms. This means that expected expenditures must be increased by inflation so that the amount budgeted will have the same purchasing power as it would have today. Similarly, revenues from CPP, OAS, GIS and any pension that is fully indexed for inflation must be increased by expected inflation so as to estimate the nominal dollars that will be available. Again, we need an estimate of the expected inflation rate.
Estimating future inflation
How can we estimate expected inflation over periods of one to thirty years into the future? Three main approaches are available:
- One can look at the opinions of professional economists and others and derive an average or consensus view. There are many sources for such estimates. One widely-available source is by The Economist magazine, which publishes inflation forecasts for the current and next calendar years, updated once a month. However, such forecasts are generally very short run and not very useful for longer term planning.
- One can compare rates of return on nominal bonds and real Real Return Bonds (RRBs) in Canada or nominal bonds and treasury inflation protected bonds (TIPs) in the U.S., taking care that the bonds being compared are similar in maturity, premium, and other aspects. The difference between the nominal rate (on the nominal bonds) and the real rate (on the real-return bonds) for the maturity being examined is made up of three components: (a) expected inflation over the common maturity of the bonds (b) a risk premium since the bond-holder has transferred any risk of unanticipated inflation from himself to the issuer of the bond (c) a liquidity premium, because the volume of nominal bonds is usally much greater than the volume of real bonds and so the former are more liquid. The liquidity premium is usally very small and can be ignored (although it did spike to some 150 basis points during the 2008 financial crisis). The risk premium for unanticipated inflation also varies over time, but is usually between 50 and 80 basis points. The expected rate of return can then be calculated. For example, as of 6 December 2012, the Bank of Canada showed the nominal return on 30-year nominal Government of Canada bonds to be 2.30% and the real rate on similar real bonds to be 0.33%. The gap is 1.97%. Assuming a liquidity premium of 0.08% and an unanticipated inflation risk premium of 0.60%, the anticipated inflation rate is 1.93% + 0.08% - 0.60% = 1.41%, i.e. the anticipated inflation rate over the next thirty years is 1.4%.
- The Bank of Canada has issued guidelines to the effect that it intends to maintain long-term inflation in Canada in a band between 1% and 3%, and further that intends to target the 2% midpoint. Many planners use the 2% target as an estimate for expected inflation.
In practical terms, whatever expected inflation rate one chooses for financial planning purposes, it is wise to calculate the results if realized inflation is, say, 1% higher or 2% higher. Caution suggests at least looking at even higher rates, at least as high as 5%.
Deflation is a decrease over time of the general price level. Most of the discussion of deflation is generally parallel to that of inflation, except with negative rates instead of positive rates. However, there are a number of significant differences.
First, deflation usually occurs during a significant economic downturn. The most famous example is the Great Depression of the 1930s, and Japan in the 1990s and 2000s. Deflation is generally believed to worsen economic downturns, because consumers will be reluctant to buy products now, waiting until later when nominal prices will have dropped.
Canada has not suffered a significant deflation since the 1930s. Even during the 2008-09 financial crisis, when the CPI did drop for a couple of quarters, the dip was relatively small and had no lasting impact. Nevertheless, deflation is a possibility albeit a very small one. Financial planning for deflation can use the same tools as for inflation, although obviously the choice of investment assets will be very different than in an inflationary situation.
- Bank of Canada, Consumer Price Index (pdf). Viewed Dec. 12, 2012.
- Consumer Price Index, historical summary. Viewed December 12, 2012.
- The Consumer Price Index: Data quality, concepts and methodology: Data quality, concepts and methodology. Viewed December 12, 2012.
- Inflation-Control Target - Bank of Canada. Viewed December 11, 2012.