Blog: 16 February 2017 (Reserve Bank High-Low Inflation forecast range and where inflation came in)
|
Related Topics |
In today's blog we take a look at South African Reserve Bank (SARB)'s High-Low forecasts for inflation for various quarters and then compare where the actual inflation numbers came in at. As we have discussed in the past, the forecasting of inflation by SARB is not very good, and this has serious implications for South Africa's economy. As poorly forecasted inflation leads to poor decisions made by SARB monetary policy committee which can (and has we might add) lead to incorrect decisions made on interest rates. For example raising rates due to expected inflation being outside the 3% to 6% range (and that expected inflation is based on poor inflation forecasting)
|
Wide range of values over short period of time
The bar chart below shows both the number of credit card transactions (the blue bars and left hand axis), as well as the total value of all of these transactions (red bars and right hand axis). While they do largely move together this only tells half the story. As the value of transactions is not adjusted for inflation. Therefore it shows nominal prices instead of real prices. In order to gain a deeper understanding of the underlying trend in credit card spending, one needs to look at value spent after adjusting for inflation. In order to determine of the growth in spending on credit card's is underlying growth or growth fuelled purely by inflation.
So how did we calculate the High-Low range? We used SARB forecasted values for specific quarters from all MPC meetings from January 2016. And compared the forecasts for each quarter of all the meetings that took place. Then ranked the forecasts per quarter from high to low. The highest and lowest value of inflation forecasted by SARB for each quarter was then used to create the blue area chart we see.
The forecasted inflation for Q1 and Q2 of 2016 was pretty tight (I.e difference between low and high value is small). With the difference between the high and low values being 0.2% and 0.3% for these quarters. However when it came to forecasting Q3 and Q4 of 2016 it seems like SARB had a really hard time in getting the forecast right for these two time periods. As at one of the MPC meetings inflation for Q3 was forecasted to be at 6.5% for the quarter, and at another meeting that forecast for the same quarter was sitting at 7% (0.5% range). Now sure 0.5% difference doesn't sound like a lot but 0.5% over 6.5% is a 7.7% difference in the forecasted value for the same quarter. Now imagine an entity like Stats SA incorrectly estimating the size of the economy by 7.7% or the unemployment rate by 7.7%. There would be mayhem.
Yet who holds the bank accountant-able when their forecasting leaves much to be desired and those forecasts ultimately have a impact on the MPC's interest rate setting policy?
But it gets even worse. At two seperate MPC meetings the forecasted inflation for Q4:2016 was 6.4% and 7.8% (a range of 1.4%). Now 1.4% range on 6.4% gives a difference 21.8%. So at two seperate meetings with the same staff at the MPC and assuming same people supplying inputs, for the same forecast period (Q4:2016), SARB derived estimates that differed by over 20%. Either their assumptions changed dramatically or their forecasters have no clue how to do their job.
Yet who holds the bank accountant-able when their forecasting leaves much to be desired and those forecasts ultimately have a impact on the MPC's interest rate setting policy?
But it gets even worse. At two seperate MPC meetings the forecasted inflation for Q4:2016 was 6.4% and 7.8% (a range of 1.4%). Now 1.4% range on 6.4% gives a difference 21.8%. So at two seperate meetings with the same staff at the MPC and assuming same people supplying inputs, for the same forecast period (Q4:2016), SARB derived estimates that differed by over 20%. Either their assumptions changed dramatically or their forecasters have no clue how to do their job.
And of course the story gets even worse. At two separate meetings the MPC forecasted inflation for the first quarter of 2017 (Q1:2017) to be either 6.1% of 7.8% (thats a range of 1.7%). And 1.7% on 6.1% gives a difference of 27.9%. That's almost a 30% difference in opinion as to where inflation will be at for the same time period. Imagine Statistics South Africa all of a sudden coming out and saying that oh we thought at this time the size of the economy was R3trillion. We made a mistake it's actually over R4trillion. Imagine the chaos that would erupt from that as policy setting is based on information supplied by Stats SA. Yet how is the forecast of inflation any different? Monetary policy setting is largely based on expected inflation which is basically their forecasted inflation.
How does the MPC make informed decisions on interest rates when their forecasts obviously cannot assist in making informed decisions as the difference in inflation estimate for the same time period is as big as 27.9%
In our predictions for 2017 and hopes and dreams for 2017 as published on Biznews we mentioned that we hope SARB gets a new team of forecasters so that we can be sure monetary policy is set based on quality inputs (including better forecasting) as the impact of using poor forecasting as a guide to inflation has significant effects on South Africa's interest rates and ultimately it's economy.
How does the MPC make informed decisions on interest rates when their forecasts obviously cannot assist in making informed decisions as the difference in inflation estimate for the same time period is as big as 27.9%
In our predictions for 2017 and hopes and dreams for 2017 as published on Biznews we mentioned that we hope SARB gets a new team of forecasters so that we can be sure monetary policy is set based on quality inputs (including better forecasting) as the impact of using poor forecasting as a guide to inflation has significant effects on South Africa's interest rates and ultimately it's economy.