The switch in the formula used to calculate monthly and annual inflation rates to take into account the proportions of foreign and local currency used to make purchases does indeed, as the authorities insist, make better sense when it comes to calculating the national inflation rate but makes it ever more difficult to work out, sometimes even approximately, what the inflation rate is for any particular individual.
The mechanism for calculating this official rate, and the same blended rate since 2020, has been explained by ZimStat. Every month, over little more than a week from roughly the end of the first week of the month, the ZimStat pricing team goes out and records several thousand prices across a range of retail outlets plus schools and other service providers, for a list of several hundred items. The prices for more than two years have been recorded in both US dollars and local currency.
ZimStat used detailed consumer spending surveys to divide consumer spending into 12 brackets, each one weighted to give a percentage of monthly consumer spending. So food and non-alcoholic beverages, easily the largest, absorbs 31,3 percent of average consumer spending. Housing, water, electricity, gas and other utilities comes in second at 27,62 percent. There is a major drop to third-placed transport on 8,39 percent.
These three categories, all fairly constant monthly expenses with spending month in and month out, total 67,31 percent, and alcohol and tobacco add another 4,90 percent, communication 2,65 percent and recreation and culture on 2,27 percent take this recurrent monthly spending to 77,13 percent of total spending.
The other six categories are, or can be, less recurrent at the monthly level. Health, at 1,42 percent will be close to monthly recurrent for those with very good medical aid or whose health spending is almost entirely fairly cheap chronic medication. But generally it will oscillate a lot between months. Education, at 4,25 percent, will be regular spending, but the payments only made every four or six months.
Clothing and footwear at 4,35 percent and furniture and equipment at 5,29 percent can be categorised as capital spending, and so very intermittent, with a lot spent at bonus time or certainly in months when school fees are not due, except some uniforms, or when someone in a household is very ill.
Restaurants and hotels at 1,08 percent seem to be made up of the odd bit of spending divided among the twelve months of the year. The grab bag of miscellaneous items at 6,46 percent could be anything, from plumbing repairs upwards.
Within each category, there will be further weighted averages, for example of what percentage of food spending goes on cooking oil or beef, and there will be average prices generated for each of those items from the thousands of monthly readings. ZimStat obviously have some sort of master formula that once the data is punched in will generate the 12 category indexes and then, the food and non-food indexes and finally the total index.
It must be stressed that this calculation and the indices it generates does not fit any one particular person. It is a triple average, first of the item prices and then two weighted averages for how each average item price affects the category spend, and then further weighted calculations on how each category index affects the final index. It can be useful, but is not specific to a particular household.
There is another set of variation largely generated by where in a socio-economic class a household lies. Generally as household income increases the amount spend on food also rises, but normally as at a lower rate than income, which tends to reduce the percentage of spending on food as spending rises. So someone at the very bottom of the ladder might spend more than half their money on food, while a very rich person might spend well under 20 or even 15 percent on food, despite eating the best cuts of steak and imported luxuries. The 31,3 percent on food used by ZimStat implies someone at very much the lower end of the middle-income socio-economic class, or even lower.
What is being done now, from the publication of the February inflation figures and cost-of-living indexes is that ZimStat punch in the US dollar and Zimbabwe dollar prices for all their listed items and generate the indices in for each category, and then overall, for each currency before then using the weighted average of these two sets of indices depending on the proportion used in total transactions for each currency to generate the new official indices and inflation rates. The indices and inflation rates for each currency cannot now be published, in terms of an instruction from the Ministry of Finance and Economic Development, based on a recommendation of the Reserve Bank of Zimbabwe in its 2023 monetary policy statement.
There is the question of how the percentage of currencies used is worked out, and at the moment we are given figures of between 70 percent and 75 percent. In theory the Reserve Bank could get an approximation from what is banked each day, but the more accurate figure would come from the Finance Ministry on the total VAT receipts in each currency.
Zimra do check, extensively, on how books are kept in each business and insist on those fiscal tills. While business details are kept private, Zimra does remit its total VAT collections to the Consolidated Revenue Fund, and so the Government and so the Reserve Bank know what the monthly totals are, and so can work out the gross percentages of VAT income in each currency. We hope this is how this is done for accuracy.
Since ZimStat have been creating a blended inflation rate since 2020, which allows like to be compared to like, it is fairly obvious that the percentage of transactions in US dollars has been rising, especially in the last three or four months, when we look at the older CPI figures in local currency, check the US inflation rates, and do a rough calculation. A year ago well over half of transactions were in local currency, and probably well over three quarters.
The rapid rise in US dollar transactions appears to date from after the taming of the black market in the middle of last year. Before people with foreign currency would pass through the black market before shopping; after that people spent their US dollars at the tills, so the blended rate makes sense since that changeover.
It is still possible to get an approximation of the Zimbabwe dollar CPI figures since the end of last year although the Reserve Bank gets these and will be using local currency inflation to set interest rates but will not allow these to be issued.
There is strong pressure, which the authorities are resisting, to give out the US dollar and local currency inflation rates along with the official blended rate, but since the US now has around 6,5 percent annual inflation, about 0,6 percent a month, a back calculation can be done which is likely to be more accurate than any other implied local currency rate which can be generated; some of us remember the complex indices used at one stage to find an approximate real exchange rate.
The biggest single problem is having a sole blended rate of inflation is that there is so much range in the percentages of currency in both earnings and spending. Hardly anyone is on a 75-25 or 70-30 split. There are many who earnings are entirely in local currency, and need, when we look at pay rises and the like, something that is more useful than the blended rate. Admittedly, if both currency rates were very low, this would not be a problem, but people would need to be convinced.
But, as we have shown any of our inflation rates are the result of at three sets of averaging and weighted averaging, and adding a fourth just moves it one step further from any actual household. But that division between the local currency households and the foreign currency households can be large. Even at the other end of the market, ZSE prices and yields are in local currency, and investors need inflation figures to work out if they are making capital gains in real terms, if dividends are rising or falling, again in real terms, and generally if their investment makes sense.
On the other hand, we have the particular Zimbabwean problem that far too many people assume that the annual inflation rate describes more than how the cost of living rose in the last 12 months, and that this annual rate describes what is happening now, and can be used predictively.
In 2020 and last year, the annual rate was of very little use. In both years, if you graph monthly inflation, which measures the rise in month CPI, you had a very rapid rise to a peak, followed by a sudden collapse, followed by a straight line tending up in the last few months of 2020 and down in last few months of 2022.
Even graphing the CPI would give a rapid upward line, followed by a few months of slow rise, followed by months of little change. At least three mathematical formulas would be needed to describe the changes over either year, so the figures were pure description of past events, with zero use, in either the upwards swing or the peaking or the downwards swing for any prescriptions. The blended rate are least reflects the reality of the total spending, although not individual spending, but then the CPI indices have never done that, either.