After years of remaining a theoretical plaything for money nerds like me, nominal GDP targeting has suddenly become a thing.   NGDP targeting has been getting love from economists across the political spectrum, which naturally means laypeople across the spectrum are responding with skepticism. Drudged up from what I remember from Prof. Burdekin's Money and Banking, I present the basics of NGDP targeting:

What is NGDP targeting?

NGDP targeting means that the central bank sets monetary policy to target nominal gross domestic product, that is, GDP that is not adjusted for price level.  The beauty of NGDP targeting is that since both prices and output are part of NGDP, both of these things influence monetary policy even though only one metric is considered.  Either NGDP level, or growth, can be targeted.

Why liberal economists like it: liberal economists tend to be more supportive of considering economic growth, not just inflation, when setting monetary policy.  NGDP targeting naturally includes growth as an input.

Why conservative economists like it: conservative economists tend to prefer simple, transparent rules for monetary policy, as opposed to opaque systems that give broad discretion to policy makers. NGDP targeting is easy to understand and easy to see if policy makers have met their goal.

So how is it different?

The Federal Reserve has what’s called a dual mandate—they are charged with considering both prices and economic output when setting policy.  (Some central banks, like the European Central Bank, are charged with only considering prices.)  The Federal Reserve does not commit to following any particular formula for setting monetary policy, but it is largely believed to behave as if it follows the Taylor Rule, meaning it considers deviations from target inflation rates and target employment rates.   Targeting NGDP would provide the Federal Reserve with a mandate to specifically target deviations from output targets.

There are other things a central bank can consider besides prices, employment, and output when setting policy, such as money growth and asset prices.  A central bank with broad discretion can consider all of these things, but this comes at the cost of making monetary policy decisions less transparent.

But what about QE2??

So where do things like quantitative easing fit into all this? It is important to distinguish between the tools used to set monetary policy and the tools used to implement monetary policy.  Tools used to set monetary policy, such as Taylor Rules, money growth target, or NGDP targeting, tell the central bank whether they need to tighten or loosen policy.  The tools used to implement monetary policy generally remain the same regardless of what policy prescription tool you use.  If interest rates are at zero, and inflation is low and unemployment high, you need quantitative easing to get to your targets.  If interest rates are at zero and NGDP is below target—guess what, you still need quantitative easing to get to your target.

Targeting Growth vs. Level

If you are going to target NGDP, you have to decide whether to target NGDP growth rate or NGDP level. Most of the cool kids in economics prefer level targeting.  Level targeting has the benefit of allowing for catch up periods: after the economy has had a downturn, it natural that it have a period of both higher growth and inflation to catch up to where it would have been pre-downturn.

So let’s do it?

To me, whether NGDP targeting should be implemented actually involves three different questions:

1.       Should GDP be considered when setting monetary policy?  The downside to considering GDP or other factors is that the central bank places less weight on price stability, which is bad if you are an inflation hawk.  The benefit is that drops in GDP often lead drops in inflation or unemployment, so considering GDP can help the central bank adjust to new economic trends more quickly.

2.       Should the central bank target levels rather than growth rates? When it comes to prices, levels are arbitrary—it’s change in prices that matter. That’s the primary logic behind targeting inflation rather than price level.  However, as discussed above, it is natural for an economy that has been sluggish for a while to have a period of higher growth and higher inflation to “catch up” after a downturn, so monetary policy that accommodates this can make sense. This accommodation can actually help end a downturn more quickly—if during a downturn, people know they can expect extra high inflation in the future, they have an incentive to spend more now, stimulating the economy.  The flip side to all this, which I haven’t seen discussed much, is that monetary policy lags the trends a bit.  This sounds great when you are coming out of a recession, but what about when you are coming out of an expansion? Would the people keen to see accommodating monetary policy continue as growth picks up be equally keen to see tight monetary policy stay in place when growth comes to an end?

3.       Should the central bank have a clearly defined rule? The benefit of a rule is that it provides transparency and sets expectations, and expectations are essential for effective monetary policy.  The downside of a rule is less flexibility, and loss of credibility if the central bank fails to meet its targets.

If you answered yes to all of the above, then NGDP level targeting is for you.  Personally, I agree that a central bank with a dual mandate should consider GDP.  However, I think that the central bank should have some flexibility in targeting levels versus growth, and that growth is sometimes the more appropriate metric.   Moreover, I think that a central bank like the Federal Reserve, that has a high level of trust and credibility, has more to lose from committing to a rule like NGDP level targeting than it has to gain, since failure to achieve the target would hurt its credibility.  As it is, the Federal Reserve can consider the policy prescriptions of an NGDP target, while also considering Taylor Rules, asset prices, and other economic data when setting monetary policy.

NPR recently featured a segment on tipping, positing that while many people believe they tip to reward good service, they actually tip out of guilt for being served by another.  The segment points out that people tend to tip at fairly constant rates, regardless of how good the service is, and more interestingly, the services that conventionally require tips in the United States are those where the person receiving the service is having a lot more fun than the server.  People at restaurants and hotels tip; people at the dentist do not.

Guilt seems to make up a large share of my (admittedly under-average) emotional spectrum, so I find this very compelling.  The segment points out a downside to this: tipping out of guilt may not be efficient.  Tippers may give an amount larger than the value of the service to them, and tips that don’t vary with service quality don’t provide incentives for better service.

Tipping norms in Ghana are quite different from those in the United States: tips are not necessarily expected for restaurant service, but for help with directions, with making a large purchase, or with loading bags onto a bus, a tip, or “dash” is expected.  (“Dash” functions as both a noun and a verb.) Often, the services you are expected to dash for are services you don’t even want, and the tipper may even give money just to get someone to go away.  The role of guilt in tipping is compounded in these situations by the uncertainty foreigners may have regarding tipping norms and by the income disparity between the average foreigner and the people who do these types of jobs in Ghana.  

The trouble with this is that the efficiency of the tip is further diminished.  While guilt may make me a generally good tipper, as an economist, I also feel guilty when I tip for poor service or services I don’t want, thus providing poor incentives.  Here is my advice on tipping in Ghana to maximize good incentives:

Restaurants: Tipping is not mandatory for food service, though it becomes more expected the more upscale the venue.  I highly recommend giving small tips/dashes that are very sensitive to the quality of service.  At a local eating spot, 1 GHS is a good tip for basic service, and will likely get you a little extra attention next time you visit.  At a nice restaurant in Accra, a 10% tip for good service seems to be well-received.  The rarity of tipping in Ghanaian restaurants presents an opportunity to tie tipping to good service, so I would urge varying tips accordingly, giving nothing for poor service, and large tips for good service.

 Bags: If someone helps you carry your bag, it is very much expected you will give a dash. If you don’t want to, then be firm about carrying your own bags. 

Bus baggage: Small dashes are often expected for loading your bags under a bus.  It is hard to avoid using this service.  This is the context where I have found demands for dashes to be most outrageous.  A dash should not be mandatory for this—I have seen supervisors yell at men who demanded a dash before loading bags. I would recommend resisting anyone who demands one before helping you.  The dash should also not be large.  I once encountered a man who refused my 1 GHS dash and demanded 2 GHS.  I gave him nothing.  I would not give more than 1 GHS unless your bags are many, or you receive some special assistance with them.  Note that a dash for loading should not be confused with an actual fee for baggage.

Directions: You should not have to pay a dash for directions.  If someone walks a long way with you to show you where something is, a dash may or may not be demanded.  If you don’t want to give one, don’t accept the escort.

Assistance with purchases: If someone helps you locate, select, bargain for, and complete a substantial purchase, a dash may be in order.  Things to consider: how much help you received, how much time and money you saved as a result of the help, and whether the person got any financial gain from your purchase.  Generally, anyone who approaches YOU about buying something doesn’t need a dash.

Household errands: If you have a guard or groundskeeper, the person will often run errands for you. You should dash to compensate for their travel costs and efforts.

Professional services: Professional services outside a person’s normal job may require a dash, depending on the job and the organization a person works with.  As an example, I had a document notarized by a judge in Tamale, and I paid a 5 GHS dash for his time and trouble.

Stealth window washing: One of the strangest things about Accra are stealth window washers, who swoop in on a car waiting at an intersection, squeegee the front windshield despite the driver’s protests, and then demand a dash after.  No matter how guilty you feel, please, please do not dash for this or other unwanted services that are forced on you; you will only encourage the practice. 

Further advice on tipping in Ghana (or anywhere else)? Please contribute in the comments!

Chris Blattman’s blog recently critiqued an article by Dan Pallotta arguing that earmarking funds for programs with proven impact is actually less impactful than using the money for further fund-raising efforts.  Pallotta makes an argument that spending on fund-raising allows you to, in essence, leverage your funds and get a much higher return on investment than you would if you’d spent that money directly on programs.

Blattman makes two counter points:

1.        The effectiveness of the programs you are funding feeds back into your ability to use your money to raise more funds.

2.       It’s not clear that lack of funds is the binding constraint in aid.

I’m  a bit skeptical of Blattman’s second point—I thought I was out here getting malaria to make sure that scarce development resources were spent on programs with the highest impact.  I think it is more correct to think of funds and good practice as being similar to labor and capital—in most circumstances you can add more of one or the other and improve outcomes, but are most effective when increased together.

I think Blattman’s first point is completely correct. Pallotta is right that fund-raising can increase impact, but program impact is fundamental to fund-raising effectiveness and meaning.  Donors should be attracted by good programs.  In a rational world with perfect information, donors would know exactly how much money they wanted to spend, and they would choose the program with the highest impact-per-dollar.  This is how these institutional funders Pallotta is complaining about behave.   However, in the real world, human behavior is less rational and more suggestible.  If fund-raising can actually increase the number of dollars out there to be used on development, it can indeed be highly impactful.  Note that fund-raising that just diverts funding from one project to another from a fixed pool of resources doesn’t get to claim this—unless the program it diverts money to is more impactful that the program it diverts money from.

Which brings us to the next point-- if your programs don’t have impact, it doesn’t matter how much you leverage your dollars- you are just using more money badly.  Palotta’s proposal to use seed money to fundraise is similar to the concept of hedge funds.  Hedge funds can’t make huge returns without leveraging their initial funds with loans, but if they don’t put the leveraged funds in investments with good returns, they are just wasting everyone’s money.  

Palotta also argues that you often can’t know what is going to be impactful ex ante.  That may be true, but that doesn’t mean you should throw in the towel and give up on trying to target impactful programs.   Market investors often can’t know which stocks will take off, but no investor would throw money at one without trying to make an educated assessment of its future value.   If funding truly is a scarce resource, you have to have some standard for choosing which programs to fund and which not to. 

Polatta may be right to encourage donors to allow their funds to be used for further fundraising, but this only makes sense in concert with an emphasis on evaluation.  After all, what is the point of all that fundraising if you aren’t going to do anything good with it?  And for fundraising to matter, Blattman must be wrong about money not being a binding constraint.  If money is a binding constraint, then you can’t fund everything, and it becomes all the more crucial to have some way of assessing the best programs.