I usually don’t comment when an economist responds to what I am supposed to have said in press quotes because it leads to more confusion than clarity. However, Professor Paul Krugman has a large following, and he has criticized me repeatedly for my stance on Fed policy, especially my exhortation for the Fed to exit its policy of ultra-low interest rates sooner rather than later. I will try and address his latest concerns (http://krugman.blogs.nytimes.com/2010/08/23/making-it-up/), which are in response to a Bloomberg article (http://www.bloomberg.com/news/2010-08-22/bernanke-must-raise-benchmark-2-points-in-prescient-rajan-s-latest-warning.html). But before I start, let me declare that I am as much for getting out of this recession as anyone else. However, I would like to get out of this recession in a way that is sustainable, that does not merely pump up growth in the short term only to see it collapse later. This is where Krugman and I differ the most.
Krugman (and many others) have in mind a Keynesian world where some distortion is keeping aggregate demand unnaturally low. The favorite Keynesian explanation is sticky wages. Because corporations cannot reduce wages to the level commensurate with the (now low) value added by additional labor, corporations will not hire. Hence the Keynesian answer to the problem of continued unemployment seems to be to give corporations an interest rate subsidy to offset the additional burden caused by excessively high wages. Hopefully, this will make investment profitable, and induce corporations to invest and create jobs.
What is the appropriate level of this subsidy? Whatever it takes to bring jobs back according to Krugman. We have had short term interest rates at negative real levels for two years (which means anyone invested in money market funds gets less back than they put in). Whether this has had sizeable beneficial effect on corporate investment and employment is anyone’s guess, but it certainly has not brought unemployment down, so Krugman wants more. Most people cannot see anything wrong with this, so let me try an analogy to make my point.
Since the problem is of getting more demand, any subsidy applied to a general input should work. So following the Keynesian/Krugman model, why not subsidize energy – for example, cut oil, gas, and coal prices substantially and keep them low “for an extended period”? For corporations will see their energy bills cut substantially, and see profits rise, so they can employ more people. Unemployment will come down, and we will move out of recession. Eventually, we can withdraw the subsidy when the economy is healthier. The benefits are clear. We will get out of the recession and put many unfortunate people back to work. And the sooner we put people back to work, the less long term damage is done to their employability. So why not do it now?
The sensible economist will, of course, have some objections.
Reason 1: The cost. Someone has to pay for the energy subsidy. If the government simply declares fuel prices to be half of what they currently are and enforces the lower price, fuel suppliers will shut down and refuse to sell anything. So the government has to pay the subsidy upfront to suppliers for them to supply at the mandated price. Even if the government absorbs the cost today, the taxpayer will have to fork out sometime. So the obvious and immediate cost will make taxpayers (and the government) think carefully about whether they want fuel subsidies.
Why cannot the same argument be made for the subsidy implicit in low rates? After all, there is a direct cost to maintaining ultra-low rates, and it is paid by anyone who has financial assets and is forgoing the normal return on them. Assume owners of financial assets hold 20 trillion dollars worth of them (this is an underestimate) and are paid 2% lower real rates per year than they would otherwise obtain because of the Fed’s policy of maintaining zero nominal rates. This implies an annual stimulus of $400 billion in real money, off the backs of those who own financial assets. Given that it has been going on for two years, this is a subsidy that is now bigger than TARP (and coincidentally, also goes largely to bankers who are the biggest short term borrowers in the market). If the government raised taxes explicitly to provide the interest subsidy, everyone would scrutinize the use this money was being put to carefully. Because the Fed picks investors’ pockets silently and forcibly through its ability to set the short term interest rate, no one asks questions about cost.
Reason 2. The short term distortions. A second reason not to subsidize fuel is that it will distort activity. Drivers who are environmentally conscious will be penalized while those with Hummers will be rewarded, with consequent damage to the environment. The biggest subsidies will go to those who are most irresponsible about their energy use. This will keep the economy from becoming more energy efficient, and make it uncompetitive in the longer run.
Again similar arguments can be made about ultra-low interest rates. Ultra-low rates encourage people to borrow to acquire assets, and are partly responsible for both the over-building in housing, the over-indebtedness of households, as well as the over-leveraging of the financial sector. More generally, a subsidy to capital will imply greater capital intensity (and waste) of capital, greater short term leverage, and excessive growth of sectors that rely on either fixed asset investment or credit. Is this the appropriate way to go (especially if we want more labor intensive sectors to grow to provide the jobs that are needed), and is it sustainable?
Reason 3. The long term distortions. Suppose every time the population splurged on Hummers, and caused energy prices to go through the roof, the government, seeing the damage to the economy, came through with its energy subsidy scheme. The government would justify its actions saying it was not willing to see the economy crash simply to uphold a vague theoretical ivory tower principle such as “moral hazard”. Such actions would reinforce the incentives of people to buy Hummers (provided they could see enough others doing so), and reduce the incentives for people to buy a Prius. In other words, the prospect of periodic subsidies could create the very actions that bring about the need for subsidies.
The analogy to the moral hazard that is engendered in financial markets should be clear: The asymmetric Fed policy of cutting rates sharply when the economy is in trouble and not raising them quickly when it recovers gives the financial sector lower incentives to worry about credit or liquidity risk. The financial sector has come to rely on the Fed to bail it out through ultra-low interest rates whenever it gets into trouble and the Fed has developed a reputation of obliging. Indeed, the Fed is now trapped because of the expectations it has set -- because the market “expects” ultra-low rates, the Fed cannot even return to normal low rates without the market taking fright. And it is hard to find a Wall Street economist who is not urging the Fed to undertake stronger, unorthodox actions.
I don’t want to push the analogy too far. But you get the point – cutting rates is not without cost. But what about the benefits? Are they as large as the Keynesians state them to be?
The real problem is corporations are not hiring quickly. But corporations did not hire quickly following the recession of 2001 (or that of 1990-91), and the sustained monetary stimulus that many economists supported then led, in no small part, to the housing boom and bust. It did not, however, lead to an explosion in corporate investment. Before saying the real problem is we are not providing enough monetary stimulus, should we not worry about why corporations did not invest then and what other problems will emerge as we keep rates ultra-low while hoping corporations will see the light? I am not arguing that ultra-low interest rates will have no effect on investment, only that I am not convinced the effect (relative to merely low interest rates) is huge, and recent history bears me out.
I would rather that more emphasis be placed on improving skill levels in the unemployed work force, and assisting the unemployed to both prepare for, and look for, the jobs that are being created in the economy. This will take time, but probably be more effective than praying ultra-low rates bring back the jobs.
Before concluding, let me emphasize that even though I discussed the direct costs of ultra-low rates above, my greater concern is with indirect costs. The way low interest rates work (apart from the direct cost of capital effect), is by raising asset prices and incentivizing investment in riskier assets. Even as corporations proved unwilling to invest last time, house prices rose, households could borrow more, and lending became increasingly crazy. Of course, bond prices aside, there are not many hints of asset bubbles in industrial countries today. But by the time a central banker stares a bubble in its face, it is too late. After a few years of rising house prices in the period 2001-2004, every broker was peddling houses as a dream investment that could never fail. So what if the Fed was raising interest rates at a measured pace from 2004? By that time, house prices were clipping along at double digits every year, and Fed rate increases made little difference. Of course, the Fed now disingenuously claims that the worst excesses in the housing market were committed when it had already started raising rates, and therefore it is not responsible for the housing boom. But it was complicit in setting off the boom by keeping interest rates too low for too long before then!
I have been making these points long enough that quotes should not be taken out of context. I am not advocating that the Fed raise rates to 2 percent overnight. That would be irresponsible. I am saying that as worries (largely about financial turmoil emanating from Europe) settle down and we return to expecting steady but slow growth, we should not wait for employment to come back substantially before we start the process of raising rates to a low normal. The Fed should, of course, give the markets a clear sense of the path to expect, and perhaps indicate it will pause after reaching the "normal" low rate. If asked what a "normal" low rate is, I would say one candidate is zero real short term rates. With inflation between 1.5 to 2 percent, that would mean 1.5 percent to 2 percent nominal before the Fed pauses again.
Post script: Professor Krugman asks if I have a model or if I am making things up as I go. He clearly knows the vast literature in banking and finance that makes some of these points about the costs of low rates. But specifically, Douglas Diamond (the father of the modern study of banking) and I, as well as Emmanuel Farhi and Jean Tirole have been writing models about the adverse consequences of sustained low Fed rates. Bill White, for whom I have great admiration for his prescient warnings when he was at the BIS during the bubble years, as well as his able colleague, Claudio Borio, have an enormous body of well-argued papers as well as empirical work warning about the effects of ultra-loose monetary policy. There are a number of papers now showing that low interest rates induce risk taking. And this is just the tip of the iceberg. Professor Krugman, there is a lot of work out there for anyone who cares to read it!