Itaú BBA - Itaú Activity Surprise Index - Data methodology changes in Brazil distort April’s result

Macro Vision

< Volver

Itaú Activity Surprise Index - Data methodology changes in Brazil distort April’s result

mayo 3, 2017

Our Itaú Activity Surprise Index inched up to to 0.12 in April from 0.11 in March

Our Itaú Activity Surprise Index inched up to to 0.12 in April from 0.11 in March, led by a methodological distortion in Brazil’s retail sales number. The tone of surprises is negative in general throughout the region. Mexico’s positive surprises are declining in magnitute but data still shows resillience to shocks, whereas Colombia returns to a rough patch, displaying similar deterioration seen in mid-2016. Results continue alluding to a sluggish pickup in activity.

The Itaú Activity Surprise Index compares trends in economic activity indicators released during the month to what analysts had been expecting for them. It is a GDP-weighted average of separate indexes for Brazil, Mexico, Chile, Colombia and Peru. An above-zero reading means favorable surprises. Below zero means disappointment. The index is a three-month average in order to avoid excess volatility. Surprises in activity often trigger revisions in GDP growth estimates.

Brazil's index registered 0.28 in April from 0.07 in the previous month. Upward revisions in the retail (and service) surveys prompted positive surprises in February: sales declined 3.2% year-over-year when the market was expecting a 7.0% decline. Had sales come in line with expectations, the aggregate LatAm index would have deteriorated. The positive surprise from the Central Bank’s monthly activity index, which incorporates these positive bumps from retail and services surveys, may have been influenced by such changes as well. Ignoring these somewhat artificial upshots, data tilted more to the negative side. Net formal job creation was negative at 64k while the market was expecting +5k. February’s industrial production also disappointed, and and its most recent March release, to be incorporated in next month’s report, came in well below estimates again. A rebound in industrial production is essential for GDP to maintain a positive trend throughout the year, following what appears to be a substantially positive 1Q17 (we forecast a 1.4% qoq/sa increase with an important contribution from agriculture). Altogether, recent data points to a slow recovery of economic activity.

Mexico's index decreased to 0.10 in April from 0.39 in March. Mexico’s positive surprises are declining in magnitute but data still shows resillience to shocks. The main highlight was from retail sales in February, surprising to the upside, up 3.6% while the market expected a 3% gain. Despite the upswing, we expect consumption, and hence retail sales, to show lower growth rates in coming months, amid a deterioration of fundamentals that supported consumers in the past years.

Chile's index improved to -0.18 in April from -0.34 in March, owing to the moving average dynamic. Whereas manufactuing, the unemployment rate, and the monthly economic activity index all yielded no surprises according to our methodology, February’s retail sales contracted 0.3% year-over-year (mkt: +0.7%). We expect private consumption to moderate ahead as the labor market (particularly waged employment growth) weakens and nominal wage growth slows. For now, we expect a 1.8% growth rate this year with a downside bias (coming from 1.6% in 2016).

Colombia's index worsened to -0.55 in April from -0.08 in March. Activity indicators for February came in notably frail and well below market expectations, consolidating activity’s weak footing at the start of the year. Retail sales set back -7.2% year-over-year, against the market’s estimate at -1.6%. The loosening labor market and depressed confidence do not hint at a swift recovery of slaes ahead. Industrial production contracted 3.2% year over year, well below the market consensus of -1.2%. Our recovery outlook for Colombia – growth picking up from 2.0% last year to 2.3% in 2017 – does include downside risks.

Peru's index declined to -0.36 in April from -0.16 in March. The monthly activity indicator for February disappointed expectations, growing only 0.7% when the market expected a growth rate of 1.8%, suggesting that activity was weak even before the economy bore the brunt of flooding and landslides caused by the “coastal El Niño”. Currently, activity is being hit by El Niño and the interruption of large infrastructure projects (stemming from a corruption scandal involving previous governments and construction firms). We believe these shocks will be partly offset by an increase in the terms of trade (after five consecutive years of decline) and the government’s fiscal stimulus package. Nevertheless, we have reduced our GDP growth forecast for 2017 to 3.3% (from 3.8%).

Find our surprise indexes on Bloomberg:

LatAm: ITMRLAI

Brazil: ITMRBI

Mexico: ITMRMI

Chile: ITMRCHLI

Colombia: ITMRCOLI

Peru: ITMRPI

Find our surprise indexes on Broadcast:

LatAm: ITSLA

Brazil: ITSBR

Mexico: ITSMX

Chile: ITSCH

Colombia: ITSCO

Peru: ITSPR

Methodology Note

Our Itaú Surprise Index LatAm compares trends in economic activity indicators to what analysts had been expecting for them each month. The index considers the month that each indicator is released. Previously, the index was built considering the month that each indicator referred to. For instance, February’s industrial production released on March will be incorporated to March’s surprise index (before: February’s index).

The index is a GDP-weighted average of separate indexes for Brazil, Mexico, Chile, Colombia and Peru. An above-zero reading means good surprises. Below zero means disappointment. The index is a three-month average in order to avoid excess volatility.

We build the surprise index for each country using all activity indicators for which consensus estimates are normally provided in the Bloomberg survey. The weight of each indicator in the index depends on its importance for the economy. For example, GDP numbers enjoy a higher weight than consumer confidence and PMIs.

We use the deviation of the actual print from the consensus estimate (surprise), subtract the result from the historical average deviation and then divide the result by the standard deviation of the surprise. This methodology provides a better sense of how important was the surprise in each month.

The weight of each country in the aggregated LatAm Surprise Index depends on the size of its GDP. Brazil has the highest weight, followed by Mexico.

It’s worth noting that, due to revisions in the economic indicators and as lagged results are published (example: GDP), the surprise indexes may be revised.

Indicators on which the index is built:

Brazil: Caged Payrolls, Unemployment Rate, Exports, Imports, Retail Sales, Industrial Production, GDP, IBC-Br monthly GDP.

Mexico: Manufacturing PMI, Service PMI, Consumer Confidence, Investment, Industrial Production, Retail Sales, IGAE monthly GDP.

Chile: Manufacturing Production, Retail Sales, Unemployment Rate, Imacec monthly GDP.

Colombia: GDP, Industrial Production, Retail Sales, Unemployment Rate.

Peru: Monthly GDP, Unemployment Rate.


 

Luka Barbosa
Lourenço Paiva


 


 



< Volver