This week, Senators Tom Coburn (R-OK) and John McCain (R-AZ) released a report critical of the federal stmulus program entitled Summertime Blues: !00 Stimulus Projects That Gave Taxpayers the Blues. It makes for depressing reading, but suffers from the very thing that garners its headlines: Anecdotes and case studies that highlights obviously worthless projects that got stimulus funding. Making broad conclusions about the the effectiveness of the federal response to the recession is suspect based on this kind of empirical analysis.
A more important report released last week by prominent econmists Alan Blinder and Mark Zandi is likely to have more staying power because of its academic rigor and it says what the Obama Adminstration had hoped all along: Federal policy ended the Great Recession and, in fact, averted another Great Depression. In essence, Blinder and Zandi (B&Z) argue that the economy has emerged from the recession and the federal government (Republicans and Democrats) can get most, if not all, the credit because of its proactive approach to fiscal and monetary policy.
Unfortunately, few people have a statistical and economic background deep enough to critique the Blinder and Zandi report. Indeed, I haven't seen much substantive criticism since the report was released save for a few speculative comments in mainstream news reports. I've now read the report and think it falls short of its grand claims.
So, with all due respect to Blinder and Zandi, I'll ask for a little patience to summarize my main criticisms of their report:
- The B&Z report relies on the Moody's Analytics model of the U.S. economy. This isn't necessarily a flaw; the Moody's model is one of the most widely used and respected ones in the business. But, in the area of macroeconomic forecasting, that's not saying much because these models are notoriously unreliable as forecasting tools. Notably, virtually all formal models failed to predict the housing bust and Great Recession, and the Moody's model was no exception. So, relying on an unreliable model to begin with should give us some pause for interpreting the results generated by the model unless major revisions are made the model and the new assumptions clarified and relationships specified. As the old computer programming saying goes: garbage in, garbage out.
- Further, the Moody's model was not subject to major revisions for the B&Z analysis. Thus, the fundamental components of the model are intact, and the changes made by the authors were little more than tweaks around the edges. Thus, the model assumes nothing fundamental has changed in the macroeconomy in terms of cause and effect. This begs the question: How can a model that failed to predict the Great Recession be relied upon to predict future economic growth, particularly during the most volatile periods of readjustment? In my view, it can't, and this may be the achille's heel of the entire analysis. No where was this issue addressed in the report.
- The model, like most macroeconomic models, relies largely on changes in spending (aggregate demand) to estimate the effects on the economy and forecast future growth. To its benefit, the Moody's model has a supply-side component, but the variables used to capture these effects are highly aggregated (using a Cobb-Douglas production function). In short, the model is fundamentally incapable of capturing dynamic supply-side shifts within sectors and industries, including changes in expectations, uncertainty, and behavioral aspects of investments decisions that are critical to triggering investment.
- Macroeconomic models, including Moody's, have no practical way of distinguishing bad investments from good investments. Literally a purely redistributive project like digging a hole and filling it up again will count the same as a project that builds a new road that reduces travel times (and improves productivity). The economic outcomes of these models are fueled by spending, and how fast people and businesses spend, regardless of whether the spending is efficient or productive. So, the multipliers (estimates of the turnover in dollars from spending) are applied equally to each project within industries based on averages for the entire industries. As critics have noted, many of the projects funded by the federal spending have been low-priority projects to begin with, so these impacts would be lower than the model would predict.
- Because of #4, most macroeconomic models have built-in biases in favor of government spending, whether its an unemployment check or a contract to dig holes and fill them back in, because in each case the spending counts as new dollars circulating in the economy, drivng up demand and economic activity. These models do not factor in the opportunity costs of redistributing money from one sector to another, or whether this redistribution is more or less productive. Again, these models are about spending, not economic growth.
- Income maintenance on food, housing, transportation and other normal household spending is treating the same way as strategic spending that fuels innovation, productivity and growth in new areas. Maintaining consumer spending is at best a stop gap measure, not a foundation for promoting long-term job and income growth (which depends on innovation, wealth creation, risk taking, and entrepreneurship).
- Entrepreneurship does not exist in these models, so this factor is irrelevant to the outcomes even though the most economists recognize that entrepreneurship--discovering new ways of productively investing resources--is a (if not the) critical factor in long-term economic growth. It's also the key reason why centrally planned economies fail and market-economies thrive.
- Even with the uncertain and unreliable nature of the macroeconomic model in the B&Z analysis, the spending side of the government stimulus was remarkably ineffective. Here, B&Z use a little rhetorical slight of hand to pump up the government role's in ending the recession by claiming that most of the government stimulus was in play when the economy bottomed out in the middle of 2009. But, the spending side of the fiscal stimulus was barely even out the door at this time. Most of the positive policy interventions (from their model) occurred in 2008 when the government responded to the financial market collapse and tax rebates were put in play.
- B&Z claim that as financial markets collapsed at the end of 2008, laissez-faire was not an option. This is more rhetorical slight of hand. The government has a monoopoly on money through the Federal Reserve and financial markets are very tightly regulated (regardless of political rhetoric that says its not). Moreover, as B&Z note correctly, public policy contributed substantially to the financial market collapse by creating regulatory uncertainty and failing to establish clear guidelines for public policy. So, yes, the government had to do something, but this isn't de facto criteria for success.
- Even with the flawed macroeconomic analysis, monetary policy proponents will certainly be encouraged by the results, since it seems clear in the B&Z analysis that the monetary policy response and the consumer-side of fiscal policy was more potent than the government spending side of the policy response.
Econometric analysis is at its best when it is teasing out factors that help us understand past behavior and changes. It is often at its worst when these techniques are used in an attempt to become a crystal ball into the future. The B&Z analysis does little to advance our knowledge about the relative effects and usefulness of the government's response to the housing market collapse (which began in 2007 and perhaps earlier), the recession (which began officially in December 2007), or the financial market collapse (in the fall of 2008). We have a long way to go before we understand the various ways markets of all sorts responded to this crisis. This will take many years to dis-entangle, and the process is surely likely to launch more than a few economists' careers.
So, for now, I remain unconvinced that the federal response to the recession has done more good than harm. I certainly do not believe we can claim this early into the process that the federal government averted another Great Depression (particularly when it had a hand in creating the Great Recession to begin with).