According to Keynesian economists, government hiring stimulates the private sector during a recession because government employees go out and spend much of their paychecks in the private sector, and extra demand is exactly what the private sector needs during a recession. But this year’s experience with census hiring contradicts that view.
Keynesians acknowledge that, someday, the private sector will pay taxes to finance the salaries and benefits of government employees, but this cost is said to be offset by the additional demand for goods and services produced in the private sector, which have those government employees as their customers.
Obama administration economists used this logic in January 2009 when they persuaded Congress to pass the American Recovery and Reinvestment Act, assuming a “multiplier”: that every 10 people hired under the law would create an additional six jobs. This is why they thought the law would have pushed the unemployment rate down close to 7 percent by now.
Of course, the unemployment rate is still close 10 percent. This might show that the legislation failed to stimulate the economy, or that the economy was in worse shape than the Obama administration originally thought, or both.
The Obama administration now says that employment would have fallen seven million if it weren’t for the stimulus, rather than the drop of four million that actually occurred from early 2009 to early 2010.
I don’t blame the Obama administration for underestimating the depth of the recession. But so far it has no way to derive its seven million estimate except by assuming that it had been right all along about the potency of the stimulus.
The law was admittedly a reaction to the state of the economy, but the 2010 census is a result of Article I of the United States Constitution, and thus provides an opportunity to measure effects of government hiring that might not be confounded by contemporaneous economic events.
The solid blue line in the chart below displays total payroll employment, inclusive of census workers and other government employees. The red line is employment apart from special census employment (I estimate special census employment as the difference between total federal nonpostal employment from its relatively constant amount outside the months February-July 2010).
If there were no extraordinary employment events this spring aside from the census, and, as Keynesians contend, hiring special census workers created jobs apart from the census, then the red line should have spiked like the blue one. In fact, the spike in the red line, if any, is pretty subtle.
If every 10 people hired for the census created an additional six jobs, then the only way to explain what actually happened is to contend, as shown in the dashed-blue line, that May and June had extraordinary and negative employment events that were offsetting the private sector benefits of census hiring and thereby obscuring an otherwise obvious spike in the red line.
Perhaps the stimulus benefits of public employment are both large and too subtle to be easily seen in the aggregate employment data or the private sector’s appreciation of public employment programs has been frustrated yet again by extraordinary negative events that miraculously coincide with true stimulus. Or maybe the Keynesian multiplier has been exaggerated.
2 comments:
I would suggest you have been tricked by his graph. Look at its scale with 129 million at the origin going up in increments of 1 million. Imagine the same graph drawn to proper scale with 0 at the origin. The movement up and down would barely be able to be registered. Even a counter-factual ‘downtick’ that the census simply countered would seem like so much ‘noise’ in the data.
A few other problems:
1. The multiplier is an iterative process. Guy rents an apartment from his job. Apartment complex gets $2500 rent. It spends $2400 on maintaince prep. Prep guy spends $2160 on helper and supplies, helper and supply company spends $1944 to wholesaler and weekly groceries etc. If money is turning over, say, every two weeks to one month (in the form of paychecks going out, bills getting paid etc) then your 1 million July stimulus jobs aren’t going to produce 600,000 multiplier jobs in July or even August but 600,000 multiplier jobs, say, August through February. That’s actually kind of consistent with the graph which shows a gentle upward slop in private employment during and post census.
2. The graph has a pretty unfortunate unit of measurement here, ‘jobs’. That’s kind of like a graph of ‘food eaten’ that counts a grape as a unit of food and a steak as a unit of food. Most of the census jobs, from what I understand, were either part time or somewhat less than full time $10 hr jobs. If the average job in the ‘normal private sector’ is something like $50K a year, 40 hours a week that works out to $24 hr. Your ‘spike’ is going to be muffled by the fact that a normal job is like 2.5 census jobs. Hence 10 census jobs producing 6 ‘census equiliviant’ jobs in the multiplier would actually show up as 2.4 private sector jobs in the graph. Consider that with the ‘spreading’ effect and you shouldn’t expect a bump.
3. Of course this exercise assumes everything else is equal on a moving ship in the middle of a storm. As census employment spiked what was the status of ARRPA’s spending? State budget spending cuts? The processing of disability applications through the VA? The dittering around extending unemployment benefits? The graph has a lot of bumps and spikes that need to be netted before you can find only the census impact.
3.1 By the way, this brings up another problem. This graph looks unusually smooth. This makes me wonder if the author used seasonally adjusted data? Seasonal adjustements can be pretty dramatic and could easily hide spikes that amount to less than 0.5% of the base numbers.
And some people do not work, because their mission - to guide others to correct for these activities. This is not the type of people power. These people are always looking for your motivation
custom paper writing
Post a Comment