Follow Paul Donovan

No confidence in confidence data?

| Posted by: Paul Donovan | Tags: Paul Donovan Weekly

Confidence surveys are treated with considerable reverence by investors. Today, surveys show high levels of confidence, signaling stronger economic activity than is reported in reality. There are five reasons confidence surveys may be flawed.

  • Surveys do not report growth, but are often presented as if they do. Media coverage of surveys is frequently inaccurate. A media report that "manufacturing expanded in Italy" is not the same as more optimism from an opinion poll surveying a sample of Italian manufacturers.
  • People cannot be bothered to fill in surveys. Electronic surveys litter email inboxes and fatigue has set in. This risks "selection bias" – people who do still bother to fill in surveys are not typical.
  • People fill in surveys incorrectly. For instance, surveys ask for the volume of export orders. When currency volatility causes the volume and value of exports to diverge, people report the value of export orders.
  • Surveys overreact. In recent years many surveys' volatility has increased relative to real economic volatility. This may be because survey responses are influenced by media reporting. The financial media has become more sensational in recent years.
  • Respondents may "game" the system. People filling in surveys know that the results are high profile; they may skew answers in the hope of impacting policy outcomes.

Surveys still impact financial markets. However, people investing strategically should give relatively low weight to surveys in their deliberations.