Table 4 Measures for valuation accuracy... | Emerald Publishing

Table 4

Measures for valuation accuracy

Statistical measure	Purpose	Recommended standard (IAAO/other)	Interpretation	Obtained	Decision
Median ASR	Central tendency of accuracy (skewed distribution/has outliers)	Between 0.90 and 1.10	A median ASR of 1 would mean a perfect (accurate) prediction. Anything <0.90 suggests under-assessment. Anything >1.10 suggests an over-assessment	0,90	Acceptable
MAPE	Measures the degree of error (inaccuracy)	Preferably below 10%, acceptable up to 20%	The properties captured in the sample are largely residential but include a few income-producing properties. Therefore, the applicable standard for the MAPE is 20%	18,98%	Acceptable
RMSE	Measures inaccuracy/accuracy	No strict threshold; lower is better	The IAAO standard does not provide a standard. However, the measure of accuracy is expressed in currencies, e.g. rands. The lower the RMSE, the more accurate the predictions. The RMSE is sensitive to outliers	R63 122,16	Acceptable

Statistical measure	Purpose	Recommended standard (IAAO/other)	Interpretation	Obtained	Decision
Median ASR	Central tendency of accuracy (skewed distribution/has outliers)	Between 0.90 and 1.10	A median ASR of 1 would mean a perfect (accurate) prediction. Anything <0.90 suggests under-assessment. Anything >1.10 suggests an over-assessment	0,90	Acceptable
MAPE	Measures the degree of error (inaccuracy)	Preferably below 10%, acceptable up to 20%	The properties captured in the sample are largely residential but include a few income-producing properties. Therefore, the applicable standard for the MAPE is 20%	18,98%	Acceptable
RMSE	Measures inaccuracy/accuracy	No strict threshold; lower is better	The IAAO standard does not provide a standard. However, the measure of accuracy is expressed in currencies, e.g. rands. The lower the RMSE, the more accurate the predictions. The RMSE is sensitive to outliers	R63 122,16	Acceptable

Source(s): Authors’ own compilation and work