INTRODUCTION
The Birches Group tracks local labour markets with reference to competitive employers found in each locality in over 140 countries
worldwide. With the significant data gathered on local compensation packages, we believe useful projections can be established concerning possible future trends in pay growth. As our indices are derived from averages of employer salaries in the local labour market, a projection based on these indices can serve as a proxy for anticipated local labour market movement. We calculate projections based on two index salary values, one based on the midpoint of the professional staff category and the second on the midpoint of the support staff category.
By default, the projections are estimated in the currency used by a majority of the employers in our survey sample to denominate pay. We refer to this as the "market practice currency." This is local currency for most countries and US dollars for those countries where the reference labour market employers utilize dollars as the basis for compensation. New: It is also possible to manually set the currency of the regression analysis to US dollars when the market practice currency is local currency and vice versa. This could be useful when there has been a recent change in the market practice currency or if you simply wish to calculate projected values in the alternate currency. Please note however, that making projections in a currency other than that which the historical market data is based on can be risky. Not only must there be a sufficient R2 value for the regression (see below) but the projection will only hold true if the relationship between the currency used in the projection and the currency on which the local labour market is based remains fairly constant. It should be noted in this context that exchange rates between currencies can be extremely volatile.
TREND PROJECTION: REGRESSION ANALYSIS
A regression analysis is performed to determine the linear (y = mx b), exponential (y = cebx), and logarithmic (y = c ln x b) functions that best approximate the observed salary data. An abbreviation in the legend of each graph designates the linear (“Lin”—dark blue), exponential (“Exp”—magenta) and logarithmic (“Log”—yellow) trends.
The function with the lowest sum of squared prediction errors with respect to the observed data (i.e. the function with the highest R2 value) is indicated in capital letters in a table at the top right of each graph. The R2 values range from 0 to 1.0. A high R2 value indicates a strong correlation between the projected values and the observed data while a low R2 value indicates a low correlation. Statisticians generally consider a projection with an R2 value of less than 0.8 to be unusable while an R2 value of 0.85 or greater is desirable.
LOCAL CURRENCY PAY INDEX
Salary averages in local currency have been converted to a pay index and are plotted for the historical period selected by the user. These values are designated as ‘Raw’ (for raw data) in the graph legends and are plotted in light blue.
To track movement of salaries in local currency for each country, salary averages corresponding to the midpoint of the support staff category on the start date of the historical range is used to establish the index base. Any changes in salary averages after that date are expressed relative to the base value. A table above each graph indicates the actual salary average in local currency, the local currency pay index, and the US dollar value of the salary value in six month increments throughout the historical date range. Dollar values are obtained by using average exchange rates for each month in question based on market rates. The same methodology is applied to track professional salaries, based on salary averages for a midpoint in the progression from entry-level professional to senior middle manager.
PROJECTED MOVEMENT
Future pay index values appearing in the table in blue print are derived from the function with the best fit to the raw data. Obtaining the projected movement with respect to the local labour market is based on the assumption that the function with the highest R2 value is representative of the local labour market. Calculating the percentage difference between the value of the function on the present date (or end date of a historical range) and the value of the function on the “project to” date therefore yields the projected movement of the local labour market between these two points in time under the assumptions made.
US DOLLAR PAY INDEX
In countries where the reference labour market employers currently utilize US dollars as the basis for compensation, salary projections are based on a dollar pay index. We run all projections using historical data in a single currency. Where the reference labour market employers paid salaries in local currency for a period during the selected historical date range but then switched to US dollars, salaries for the entire period are converted to US dollars for the purpose of running the regression. While this method allows us to present some data for each location, the projections are of lesser value as they reflect the unpredictability of currency fluctuations. (This can be reflected by lower R2 values.) A user may be able to work around this problem by selecting a shorter historical range in order to exclude the erratic raw data reflecting currency movements.
US DOLLAR VALUES OF INDEX SALARIES AND EXCHANGE RATE MOVEMENT
The Index Salaries (USD) link graphs index salaries in US dollars for the selected historical date range at the support staff midpoint and professional category midpoint. A graph showing US dollar exchange rate movement against the local currency is also available under the “exchange rate” link. The percent devaluation of the local currency vis-à-vis the US dollar is indicated in the table below the graph. This figure is negative if the local currency has appreciated against the US dollar.
NO GUARANTEE OF FUTURE PERFORMANCE
This tool is provided as a means of estimating future movement based on the assumption that the local labour market will continue to move in the same manner as has been observed in the historical data. At best, this represents an educated guess and there is no guarantee that the local labour market will move as projected.