This article was first published in the first-quarter 2012 edition of Personal Finance magazine.
One of the most common responses to media releases about consumer inflation is “There's no way that the official inflation rate is correct – my living expenses are far higher”.
Much of the time human psychology is at play, because people's confirmation biases make them focus on the expensive items they buy and ignore the items where prices are stable or rising slowly. But there are a few consumer groups whose buying power has diminished more rapidly than that of the general population over the past few years.
Statistics South Africa (StatsSA) has tried to counterbalance the one-size-fits-all approach to the overall price index by introducing other indices for specific groups. It has done this by constructing different “baskets of goods” that are representative of the buying patterns of different groups. By comparing the price increases that different groups face with the overall inflation rate, it becomes possible to see who are the winners and who are the losers of the inflation-targeting process.
The official consumer inflation statistics include price indices for different income groups and a specially constructed index for pensioners. We will examine these indices in greater detail below.
Another (indirect) way of comparing inflation across different groups is to focus on goods and services that the different groups purchase at a relatively higher (or lower) rate than the average consumer – in effect, constructing new baskets of goods with different weightings. This exercise can be useful if we want to target specific consumers on an even more disaggregated level.
Recent articles by Bruce Cameron on persfin.co.za highlight the financial difficulties facing people of retirement age. Fewer than half of all pensioners surveyed in the 2010 Sanlam Employee Benefits Benchmark retirement survey thought that they had saved enough money to retire comfortably, and only half thought they were on track for a financially secure retirement. The single biggest expense category cited in the study is that related to medical care, for which many of those surveyed said they had not saved adequately.
In 2008, StatsSA undertook to review its basket of goods and to update it. The old basket had last been revised in 2000.
This is standard practice across the world and typically involves introducing new items to the basket (such as hotel accommodation and tickets to sporting events), removing items that are no longer purchased in significant quantities (such as VHS recorders and cassettes) and changing the weightings (contributions) of other items in the basket. StatsSA went further and reviewed its entire methodology for categorising items in the basket (see the table “Weighting of categories in the 2000 and the 2008 basket of goods”, link at the end of this article). These changes in categorisation make it harder to compare certain items (or groups of items) in the new basket with those in the old basket, particularly those that affect pensioners disproportionately. As a result, all comparisons of year-on-year growth in prices look only at data from 2009 onwards. Most of the trends identified over this period were also present in the old CPI basket.
The biggest changes between the two baskets are the relatively lower weightings allocated to the “food and non-alcoholic beverages”, “health”, and “housing and utilities” categories, and the higher weightings given to the “alcoholic beverages and tobacco”, “transport” and “miscellaneous” categories. Some of the changes reflect what economists call income and substitution effects – for example, as people's income has increased they have spent less of their total income on food. Other changes are due to the abovementioned reclassification process – for example, car insurance (and other insurance payments) is now grouped under “miscellaneous” rather than with other expenditure on “transport”.
In general, pensioners are more likely to spend their money on services than on goods, and particularly on services that have shown price increases above the average rate of inflation for most of the past three years.
StatsSA has constructed separate indices for households with different income (and therefore consumption) levels.
All of the households of South Africa are divided into five groups of equal size (quintiles), organised by income levels. Roughly speaking, the richest 20 percent of households spend R9 000 or more a month, whereas the poorest 20 percent spend roughly R1 700 a month or less. The table “Different income bands' contribution to CPI” (link at the end of this article) provides estimates of the contribution to CPI by the different income bands in 2010.
Because of the enormous spending power of quintile five, its weighting in the overall headline inflation basket is 71.6 percent - in other words, the richest 20 percent of households are responsible for 71.6 percent of all the spending in the overall basket. The next richest quintile (quintile four) contributes another 15.6 percent to the overall basket. The poorest 20 percent of households are responsible for only 1.8 percent of all spending.
What this means is that the overall inflation picture most closely resembles the experiences of the top 20 percent of the population and not the bottom 60 percent. The inflation rate for the poorer quintiles can be significantly above or below the official rate.
The graph “Year-on-year inflation rates for the five income groups” (link at the end of this article) shows clearly how the inflation rates have diverged in the past few months.
For most of 2009, poorer consumers faced higher price growth than their richer neighbours, with prices in their basket increasing by more than 12 percent year-on-year. From late 2009 until the start of 2011, the situation was reversed, with high-income consumers facing the highest price growth, although this never exceeded the price growth for the poorest consumers by more than three percent.
For 2011, the situation had once again reversed, and low-income consumers faced the highest prices. In October 2011, the overall CPI rate increased by six percent. For the richest 20 percent, the increase was 5.6 percent, but for all the other groups the increase was above seven percent.
For the poorest consumers, there is a strong correlation between the growth in food prices and the growth in overall inflation, because food purchases make up a large portion of their overall consumption basket. For the richest quintile of households, food comprises only 10 percent of their consumption basket. For the next-richest quintile, the food portion of the basket jumps to 25 percent, and for the poorest quintile, food makes up 40 percent of their consumption. The trends illustrated in the graph “Headline CPI vs CPI for food and non-alcoholic beverages” (link at the end of this article) are very similar to the ones in the previous graph.
For most of 2009, and for most of 2011, the growth in the prices of food and non-alcoholic beverages easily outstripped growth in headline inflation. These periods are the same ones where the poorest consumers faced the highest growth in prices. The period from late 2009 through 2010 saw food price growth fall below the general inflation rate – exactly the same period when the poorest consumers faced the lowest price growth in their basket of goods.
With food inflation expected to remain high over the next few months, the inflation outlook for poorer households is very worrying. With many government and private policy decisions (increases in wages and social grants) benchmarked on the overall inflation figure, there is a real risk that the buying power of low-income households will be eroded in real terms if their income does not keep pace with the inflation rate they face.
StatsSA has constructed a price index specifically aimed at pensioners. The table “Headline inflation basket versus pensioners' inflation basket in 2008” (link at the end of this article) compares the weightings of the headline inflation basket with the basket of goods constructed for pensioners.
In general, the basket for pensioners most closely resembles the consumption patterns of the top two quintiles, although there are some notable exceptions. As a group, pensioners are a high-expenditure group, presumably because richer households have a much higher life expectancy than poorer ones. Pensioners spend significantly more on food and housing than do the top quintile. They spend a greater share of their income on housing and utilities than even the richest quintile does.
However, a large portion of this category is “owners' equivalent rent”, which is not an actual monthly expense; it is the estimated rent that a person would pay to rent his or her own house or one of equivalent value. Many pensioners own their homes and do not rent, and they have built up substantial equity in their homes. This item in the basket is therefore likely to overstate the inflation story for pensioners. Pensioners, similarly, pay the highest property rates and taxes.
The graph “General inflation vs pensioners' inflation” (link at the end of this article) compares the year-on-year growth in the headline CPI basket with the CPI for the pensioners' basket. It is clear that the growth in pensioners' CPI is consistently above the growth in headline CPI – in early 2009 by as much as one percent. The basket of goods that measures inflation for pensioners is an average: some pensioners may face price increases that are greater (or less) than the experience of the “average pensioner”.
In late 2009 and early 2010, the two price indices converged, but since mid 2010 they have once again diverged, with price increases for pensioners outstripping general inflation by between 0.5 percent and one percent.
The graph “Headline CPI vs health category vs medical products” (link at the end of this article) compares the growth in headline CPI with the growth in prices for the “health” category of goods and services. This category is divided into “medical products” (also shown on the graph) and “medical services” (shown on the next graph).
For most of the past three years, health-related costs have risen faster than overall prices – for most of 2009 these costs have risen by four to five percentage points more than headline CPI. During this period, the increase in the cost of medical products has been even higher, peaking at year-on-year growth of almost 16 percent in early 2009.
Towards the end of 2011, the increase in health-related costs converged on the overall inflation rate. This may in part be due to base effects; having enjoyed large price increases for most of 2009 and 2010, the medical industry may be under pressure to make smaller mark-ups for the present. There is no guarantee that this category of prices will not outstrip the general inflation rate in the future.
Health-related costs make up 1.5 percent of the headline CPI basket but 2.4 percent of the pensioners' basket.
The graph “Headline CPI vs doctors vs hospital services” (link at the end of this article) compares growth in headline CPI with growth in the prices of certain medical services (the other sub-category of the “health” category). The increase in the prices of doctors' services and hospital services is shown.
The trends are the same as in the previous graph: the growth in the prices of these services converged on the overall inflation rate in early 2009 and late 2011 but has been significantly higher for the two- and-a-half years in between. During the second half of 2010, when general inflation bottomed out at close to three percent, the inflation rate for hospital services was twice the headline rate, and the inflation rate for doctors' services was almost three times the headline rate.
The survey for doctors' services is conducted only once a year and may over-represent the actual price increases. It is possible that the growth in these prices may have started to slow to 2011 levels a few months earlier. However, it would still indicate price increases that are more than double the official inflation rate.
Medical services has a weighting of 0.9 percent in the overall CPI basket but a weighting of 1.6 percent for pensioners – almost double the share of the overall basket. Hospital services comprise only 0.2 percent of the overall basket but 0.8 percent of the pensioners' basket – four times the share. If private hospital-related cost increases outstrip general price increases, this will have a much greater effect on pensioners than on the general population.
The graph “Headline CPI vs insurance vs medical insurance” (link at the end of this article) compares growth in headline CPI with increases in the prices of general insurance and medical insurance. These services are grouped in the “miscellaneous” category.
Once again, there is convergence in price increases at the beginning and at the end of the observed periods, at least for overall insurance services. These prices still grew at more than double the rate of headline CPI from the middle of 2009 to the middle of 2010. Over the past year, these price increases have slowed dramatically, possibly as a result of greater pressure from the public.
The price increases of medical insurance, on the other hand, have been persistently high over the entire period, never dipping below 10 percent year-on-year. These numbers represent only the average increase in medical insurance rates – the increases may be even higher for those members deemed at greater risk of illness, which would include those of pensionable age.
Health insurance costs comprise 3.7 percent of the total headline CPI basket but 3.9 percent of the pensioners' basket. Only the top quintile pays more for health insurance. Remember that the different baskets measure consumption on a household level, and pensioners will have smaller households or even live alone. The costs (and price increases) on an individual level may well be higher.
The overall inflation rate (growth in headline CPI) is widely quoted and used as a reference point for wage settlements and policy targets. It is easy to forget that a large portion of the population may face price increases that are significantly higher than this rate, and that most of these people are among the most financially vulnerable in our society.
For those who are fortunate enough to be able to save and to plan for future retirement costs, it is critical that they realise that conservative savings vehicles (for example, bonds) are unlikely to offer a rate of return that will compensate for the growth in health and medical costs. These costs may rise at even faster rates in the future due to the funding requirements of national health insurance and the greater cross-subsidisation of public health users by private health users.
In order to prepare adequately for retirement, people therefore need to start saving at much younger ages and to contribute a larger percentage of their salaries than they may have previously calculated.
* Paul Berkowitz is an economist working in the field of local government and is interested in public finance, labour economics and local economic development.
PREPARE FOR HIGHER INFLATION IN RETIREMENT
You need to save more if you are to withstand the withering effects of the inflation rate you will experience in retirement.
According to Niel Gerryts, the head of the Stellenbosch branch of Alexander Forbes, most financial plans for retirement are based on the average inflation rate for everyone.
But different groups of people experience inflation in different ways, and pensioners are one such group. The average inflation rate for pensioners is higher than the overall inflation rate, which means that members of retirement funds should be saving more if they do not want to be out of pocket when they retire (see the pensioner inflation table, link at the end of this article).
Based on inflation rates between 1985 and 2011, you have to save an extra 0.5 percent of your salary, or 3.7 percent more than you are saving now, to ensure that the inflation gap in retirement does not set your financial plans askew.
And the problem has been getting worse in recent years. Taking account of the inflation difference between 2003 and 2011, the compounded difference – calculated over a projected 30 years – between average inflation and pensioner inflation has been 23.7 percent, compared with 10.7 percent for the period from 1985 to 2011.
This means retirement fund members need to increase their retirement fund contributions by 7.8 percent (or one percent of salary).
Gerryts says the impact of pensioner inflation over recent years means that investment-linked living annuity pensioners who have not saved more to take account of the higher average rate of inflation have two choices:
* They must achieve additional investment returns of 0.7 percent a year to allow for a drawdown that will match pensioner inflation instead of headline CPI. But targeting a higher return means taking on a higher level of risk.
* The initial pension must be about 7.8 percent lower to allow for increases that match pensioner inflation instead of headline CPI.
Pensioners who buy with-profit annuities, where your capital is guaranteed but your increases depend on investment returns, also need to take care, Gerryts says.
With-profit annuities should be bought at an initial purchase discount rate that is 0.7 percent lower to provide pension increases that will keep up with pensioner inflation. The higher the initial purchase discount rate, the higher your initial pension but the lower your future pension increases.
Gerryts says the trend of pensioner inflation persistently exceeding headline CPI may be expected to continue in future. The main contributors to the trend are the increasing costs of food and non-alcoholic beverages and of post-retirement medical care.
He points out that the calculation of medical inflation excludes the fact that pensioners spend more on healthcare costs as they get older. Medical inflation is based on an increase in the cost of health care only.
“It is generally good and prudent for you to plan and budget for inflation to exceed general inflation and to have additional savings to cover this risk, and to review your health and other insurance to avoid having gaps that can result in large additional costs,” Gerryts says. – Bruce Cameron