A new report from global economic research firm IHS Economics & Country Risk, backed by the Association of National Advertisers and The Advertising Coalition concludes that the nearly $300 billion in advertising expenditure in the U.S. last year drove about 16% of the nation’s sales activity, or some $5.8 trillion.
The ad spend directly stimulated about $2.4 trillion in sales, or 6.5% of the 36.7 trillion in total U.S. sales activity last year. The report surmises that 2014 ad spending drove an additional $1.4 trillion in indirect sales, the result of a “’multiplier effect’ throughout the economy as dollars flow through supply chains.”
According to the report: “The economic stimulation does not end there; companies and their suppliers hire and pay employees, who, in turn, spend some of their income in the economy on consumer goods and services. These induced consumer effects amounted to $1.7 trillion in 2014.”
Every dollar of ad spending supported, on average, about $19 of economic output (sales), per the report, while the total impact of advertising represented 19% of US GDP.
IHS forecasts ad spending rates will increase an average of 3.3% from 2014 through 2019 with total ad spend rising to nearly $350 billion.
The ANA hopes the report will help marketers persuade lawmakers that current legislative proposals to curb advertising as a business expense are ill-advised, given the positive impact the report says advertising has on the economy and the potential detrimental result if advertisers cut back because of changes in tax law.
In addition to the overall positive affect nationally that advertising creates, the report provides ad expenditure impact analyses by state and by industry.
“Advertising plays a significant role in stimulating US economic activity and supporting jobs in all sectors of the economy,” the report concludes. “Furthermore, advertising activity will continue to make a substantial contribution to the nation's economic activity through the forecast horizon, which extends to 2019.”
The full report, an executive summary and other related materials can be found here.
Aside from speculation, did the report say how this finding was definitively determined?
Ed, much of it is based on proprietary modeling; but the report has a three-page appendix devoted to Theory and Methodology. Here are a couple of passages:
"To quantify the economic impact of advertising expenditures on the US economy, this study: Estimates the total level of advertising spending in the United States and creates a 5-year forecast. Estimates sales, employment, value added and labor income impacts based on econometric models that quantify the relationship between ad spending and resulting sales. Uses input-output methodologies to compute the ripple effect of economic activity that happens as a result of the sales from ad spending. Simultaneously allocates advertising to every state, congressional district and 17 NAICS-based industry aggregates using proprietary macroeconomic, regional and industry models."
Some IRS data was used as well: "Using the IRS tax statistics database, IHS was able to collect industry-level advertising expenditure information that was reported on each corporation income tax form. A reformation of the model specification was needed as a result of revised historical data and a change in the source data of the dependent variable. The structure of the economy was much different when the model was first developed and subsequently the regressor data had a different statistical form. Thus, the old model did not provide an optimal fit of the data."
Also, there is now a link in the story to the full report.
I just waded through the report and, as I suspected, it uses various assumptions weighted with actual data to come up with its conclusion that, on average, a dollar spent in advertising generates $19 in sales. I have seen similar attempts, using various types of multiple regression models that showed much lower results, that were also questionable. The problem lies in the "intangibles"--like consumer attitudes, brand image, category elasticity, the impact of promotional activities vs. pure branding ads, etc.---which are not easily quantified and, as a rule, are not part of these models. I doubt that we will ever know, with any degree of certainty, the precise statistical relationship between ad spending and sales---except that it does exist to some extent.
Stephen, did your modeling end up with 100% attribution of the effect of all the input variables?
When I was setting up micro-econometric modelling for an Australian agency we generally found that FOR A TYPICAL BRAND we ended up with half a dozen or so significant input variables, and the total explanatory power of the model was typically in the high 60s to high 70s and occassionally into the 80s - and that is for the brands where a stable significant model could be derived. That is something like a quarter of the movement in sales couldn't be ascribed to things like price, advertising, distribution, promotion, weather, economic conditions etc.
Given your model is more of a macro-economic model it might mean that the dependent variable is more stable so you may be able to use more explantory variables and get higher degrees of fit.