When Data Cannot Be Trusted

By Paolo von Schirach–

WASHINGTON – Economics is a very imperfect science. Still, it has some shot at credibility because calculations, assumptions, assessment and scenarios worked on and presented by economists are based on the analysis of hard data. I am talking about solid numbers on the size and variations of economic sectors, public spending, GDP growth, unemployment level, tax revenue, electricity consumption, balance of trade, and a lot more.

Trusted data?

That said, what if this hard data at the foundation of economists’ calculations and, down the line, economic scenarios were not true? Imagine that some or maybe most data has been manipulated, cooked up, modified, or totally invented. Well, in that case any meaningful analysis about that particular economy is gone, because the economists would be working with fictitious numbers that have little or nothing to do with reality. There could be some useful anecdotal evidence derived from some observable phenomena, some data derived from calculations made by third parties. Still, all this would be fragmented and probably misleading or meaningless, at least in most cases.

There is manipulation

But why bring this up? Because economic data manipulation and fabrication may be more common than we think. Here is just one recent example. This is not about one particular country; but about what used to be a highly respected global index ranking countries on the easiness of doing business. I am talking about the Doing Business rankings compiled and updated every year by the supposedly neutral and intellectually unimpeachable World Bank, the large multilateral lending institution headquartered in Washington, DC.

Doing Business rankings

The Doing Business Index ranks all countries on several areas critical to assessing their “business climate“: beginning with setting up a business and then running it as smoothly as possible without the additional headaches caused by complex bureaucratic requirements and inefficient public services. For example, the Index looks at the time it takes to register a company in any given country; and then how long it takes to become operational, (purchasing land, renting space, hiring staff, getting basic services, resolve disputes, etc.). Those who compile the index look at how many procedures there are, and how long it takes to do be fully compliant. In other words, how many hurdles investors need to overcome, and a lot more. Relying on the same methodology applied to all countries, each index component gets a grade, and the country gets a final grade based on averaging the scores in all the sectors. Again, we should be reassured that everybody is treated in the same way. The same methodology is applied to all countries, big and small, rich an poor. The higher the overall score, the more business friendly and therefore investor friendly the country.

Overtime, the Doing Business rankings, updated every year, acquired a life of their own. Especially for developing countries, a good and year after year improving score was the equivalent of a “good conduct certification” provided by the most respected World Bank. A good score could be legitimately used as a marketing tool to attract foreign investors. “Come and invest here. The World Bank Doing Business Index says we are a good country, and we are getting better every year”.

But now the patina of credibility for the Doing Business index has been compromised. The World Bank suspended it in order to run an internal investigation. It seems that some countries mysteriously moved up the index ladder, without having carried out any meaningful reforms aimed at improving their domestic business climate. Others were pushed down for no clear reasons.

We do not know the facts, and therefore we should not jump to conclusions. Still, it is not impossible that some of the professionals compiling the Index may have been pressured to cook up some of the data in order to “improve” the outlook of one or more countries. There is no clear evidence about this, and therefore we have to be careful about assuming any wrong doing.

Still, this does not look good; and it may take a while before the Doing Business Index will regain its credibility and prestige. For all we know, if the integrity of the process cannot be guaranteed, the index may be discontinued.

As you can see, even the suspicion of data manipulation can have serious consequences, especially when it is about developing countries that desperately need to be taken seriously by potential investors. If confidence in the integrity of the Doing Business rankings cannot be fully restored, any investor looking at an impressive ranking improvement of any given country may conclude that this is not real, or may be somewhat unreal, and therefore not a useful tool to assess the viability of that country as an investment destination.

Argentina and Greece cooked the books

And this is not the only example. We also know that years ago official statistics in Argentina and Greece were cooked up by the government statistics agencies in charge of compiling and publishing economic data. The goal was to supply “friendly” data that would support government narratives about economic and fiscal trends. The officials in charge were ordered to present a rosy scenario, and so they did. As a result, the domestic public and the international community got a false picture. The bad data was hidden underneath fake numbers.

Chinese data

And then there is the real mystery of Chinese official economic statistics. There is no vetting, no independent calculation of anything regarding China. However, we do know that the impressive set of data coming out of the Chinese government is made public only after having been approved by the ruling Communist Party leadership. Which is to say that it is inconceivable that the Chinese government would allow the publication of data that would even minimally contradict official forecasts and goals approved by the party.

GDP always good

For instance, the Chinese Government has every interest to inflate GDP growth numbers. The power and prestige of the ruling Communist Party is based in large measure on its self-declared competence as top notch steward of the Chinese economy. Economic data that would even minimally challenge the official narrative of highly competent technocrats at the helm of the country would harm the prestige of the party.

In plain language, most of what comes out of China is molded to fit a preordained political narrative. And this means that the data released is probably untrue. Mostly false? Only a little bit false? This we cannot say for sure, simply because we have no access to the real data.

How bad is this?

Beyond these examples, how big is this problem across the world? We do not really know. In some cases, bad data can be the result of poor data collection systems and methodologies. In other words, it can be about human error, as opposed to deliberate manipulation. Still, it would be wise to assume that the problem of misleading or false numbers is more serious than we would think.

Sadly, we cannot hide from the fact that now live in a world of self-serving, often mutually exclusive, narratives. For people in power and their media and academic allies, the temptation to modify and then publicize the numbers so that they will fit their narrative is just too strong, especially in countries in which it is difficult if not impossible to independently verify what the government declares.

Problems in the US

And it gets worse. Even here in the US we have a problem. It may not be about deliberate official data manipulation. But it is about the deliberate distortions used to explain to the public what the official data really means. Taking this into account, we observe how good economic news can be presented as bad, and vice versa.

Take US GDP growth in the last couple of years, for example. Pre-Covid, it looked as if America under President Trump was doing great. Consistent GDP growth, historically low unemployment, sky high stock prices. No doubt this is data that would support the idea that, thanks to President Trump’s pro-business policies, first and foremost tax cuts and substantial deregulation affecting many economic activities, the US was powering ahead, with no end in sight.

And yet, most economists who have Democratic Party loyalties would argue that the economic trends under Trump, notwithstanding tax cuts and deregulation, did not improve that much compared to the GDP growth rate under President Obama. New business formation was unimpressive. The additional economic expansion under Trump was largely debt driven. Many US companies were alive, but not not profitable. They carried a huge debt load. They were seriously overleveraged, because they could issue bonds at almost zero interest. The stock market impressive rise was due largely to interest rate repression practiced by the US Federal Reserve. Zero interest rates made all other types of investments not attractive. Hence the concentrated focus on stocks which led to over appreciation. Finally, they would argue that at least some of this new US growth was driven by the spectacular expansion of US federal budget deficits. Indeed, pre-Covid, when everything was going super well, Uncle Sam was on course to run an annual deficit of about $ 1 trillion. This a monstrous figure in a growing economy, in peacetime, at full employment.

Healthy growth?

So, was the impressive growth under Trump real, or the result of the “steroids” provided by zero interests imposed by the Fed, and by the fiscal stimulus provided by an enormous federal budget deficit? It is really sad that in today’s America we cannot even begin to agree on any of this. In fact, we cannot even have a civil, non partisan, conversation about any of this.

Respected economists are on opposite sides of this divide, depending on their political leanings. Which is to say that, even assuming that the data everybody is looking at is genuine, we cannot agree on what it really means. And this is almost as bad as relying on cooked up figures.

Paolo von Schirach is the Editor of the Schirach Report He is also the President of the Global Policy Institute, a Washington DC think tank, and Chair of Political Sciencand International Relations at Bay Atlantic University, also in Washington, DC.




Living In A False Data World

WASHINGTON – We are used to hear that we live in a “data driven” world. Thanks to ICT and the computing power explosion, it is possible to gather, streamline and organize millions, in fact billions of bits of information. All this intelligently organized and sifted data provides precious information that will influence, in fact will determine investments, marketing strategies, purchases, and even the pitch of political campaigns.

Reliable data

Fine, all well and good. But this “data revolution” assumes that we do have reliable data; and it also assumes that we have a good sense of “what all this means”, once we have carefully looked at all the available information.

And here I see serious problems. At one level, we may engage in self-deception by giving excessive weight (or not enough weight) to some data. And, on a different level, what if the data is false, or heavily manipulated? In both cases we are in trouble.

Data in context 

Let’s talk about exaggerating the importance of some data. For example, take unemployment. The US Government is proud to tell us that at around 5% the unemployment rate is back to “normal’. True. Except that this unemployment statistic, while correct, is rarely placed in context.

Unemployment is definitely way down, and this is good news. However, the percentage of the US population that is actually working compared to the entire working age population is well below what it used to be 10 or 15 years ago. Indeed, labor participation reached 67.3% in 2000. And now in 2016 it is at down to 63%. Which really means fewer people employed. In other words, the often cited unemployment data omits the fact that millions of Americans simply dropped out of the work force, and therefore are no longer counted. Therefore, if we look at how many Americans are actually working compared to the overall working age population the picture is not so good.

Likewise, when employment statistics are presented, we usually get an aggregate number or percentage. The conventional wisdom is that if unemployment is down this must be good news. This means that the economy is growing, and therefore there is demand for more labor. What can be wrong with this? Well, nothing wrong really.

From a different angle 

However, if we look just a bit under the surface, we see that among the millions of new jobs that have been added in the last few years very few are related to productive activities (manufacturing, mining, energy) that produce new wealth. We have millions who found work in the hospitality industry, or who are on the lower echelons of the health care industry, along with many janitors and landscaping workers. Most of these are low pay, often part-time, marginal jobs.

Therefore, if you focus only on the overall unemployment figures and conclude that all is well, because a healthy labor market is a powerful indicator of a healthy economy, you are engaging in self-deception.

Over valued stock market 

On a different level, a buoyant stock market used to be considered a sign of a healthy economy. But these days it is no longer so. US share prices are dangerously inflated for reasons that have nothing to do with any misinterpretation of the real economy and market forces.

They are inflated because the US Federal Reserve continues to keep real interest rates at historically low levels. This unusual ZIRP policy created a bias against any form of saving. Since they get zero per cent keeping their money in the bank, people seeking some return on their money are induced to invest in the stock market. And since millions started buying stocks for lack of any alternatives, this has inflated stock values.

So, at this time the stock market data is not a helpful indicator of anything regarding the real economy because valuations are grossly inflated. High valuations are disconnected from economic performance. Of course, there have been bubbles before and we can expect more in the future. But this is a gigantic bubble created by the Federal Reserve and its monetary policies. The shares valuations data does not capture any of this.

False data 

And now let’s get to the bigger problem: false data. When the Greek debt crisis emerged back in 2009 it became obvious that the truth about the impending fiscal disaster had not emerged up to that point in part thanks to the complicity of the national statistics agency, (Hellenic Statistical Authority, ELSTAT). At the behest of the Greek government, the agency was happily producing false data, with the obvious intention of hiding the truth about the huge fiscal hole for as long as possible.

More recently, right after Mauricio Macri was elected President of Argentina, one of his first moves was to get new staff in the national statistics agency, (INDEC). His goal is to recreate credibility for economic data published by his new government. It is clear that the previous administration published false (or distorted) data in order to convey the message that the tottering economy was in fact doing well under their stewardship.

Just a few bad apples? 

Well, these are some of the cases we know about. But are these just a few exceptions? Are all other governments around the world complying with high ethical and professional standards when it comes to reporting economic statistics? I would not be so sure. For example, a major country in Africa, beyond inflating GDP growth statistics, cuts the actual number of its very large population in order to show a higher per capita GDP, this way trying to show a sign of economic progress that is not really there.

And then we have impeachment procedures against Dilma Rousseff, the President of Brazil, accused of manipulating public accounts in order to show a healthier fiscal situation. And what about India’s GDP numbers? Most experts argue that they are inflated, even though it is not clear by how much. In other words, India is also under suspicion of “cooking the books” in order to create a brighter economic picture.

China’s GDP numbers 

And, finally, the real monster: China’s GDP growth figures. Nobody believes the official Chinese data anymore. No, China does not grow at almost 7% a year. The question is: how big a lie is this? Is the real GDP growth 6%, or is it 3%? We simply do not know. There are many theories but no hard facts, simply because nobody trusts the official Chinese data.

Now, think about all this for a moment. China is the second largest economy in the world. And yet most experts and analysts routinely argue that the official numbers are fake. But why would China do this? It is quite simple. In China positive economic statistics are necessary tools to strengthen the regime’s political legitimacy. Inflated growth numbers tell everybody a good story: the Communist Party leadership is doing a splendid job.

What about everybody else? 

Once again, are we talking about just a few cases of rogue governments that do not play by the rules? Or is this fraudulent manipulation of sensitive economic data far more extensive?

I would say that the likelihood of data manipulation increases with the degree of authoritarianism. A government not held accountable by any one is not interested in enforcing high standards of truth and transparency. You can bet that it will say whatever it can to make itself look good. As there are not so many accountable democratic governments around the world, we can safely conclude that much of what is published and is then used by analysts as “data” is at least inaccurate, in some instances totally false.

Bad consequences

And data manipulation has really bad consequences. Unless a company enjoys the benefits of political favors, it is hard to make major economic decisions when you literally “do not know what’s going on” in any country that is in the habit of manipulating important economic data. Likewise, it is hard to attract serious foreign investors when you cannot reassure them that the country is ruled according to proper transparency standards.

Data driven world, with many lies 

So, here is the thing. We live in a very strange and paradoxical world. The IT experts tell you that they routinely capture millions of pieces of information on consumers, their preferences, their habits and buying patterns. And all this data drives decisions and investments by large corporations.

At the same time, we see how statistics, even when correct, are routinely manipulated in order to fit a preordained (and often dishonest) narrative. If you want to make the case that the US economy is doing fine, you can point to hard data: 5% unemployment, 2% GDP growth, historically high stock market valuations, low inflation.

But if you want to paint a darker picture, you will point to other hard data. 2% GDP growth is 1/3 below the historic 3% norm. On the basis of other real data, you will say that most of the new jobs are part-time gigs that at best provide survival wages, without creating any chance of upward mobility. You will argue that there are millions of part-time workers who would rather have full-time jobs but cannot get them. And you will also say that, based on hard data, (real corporate earnings for instance), the US stock market is over valued, thanks to Fed policies.

No reliable data without accountable governments 

Once again, regarding the wider world, you can rest assured that every non democratic regime in which leaders are not held accountable –and there many of them– economic numbers are either false or heavily manipulated, so that they can be used by the leaders to support a self-serving political narrative.

Yes, this is a data driven world. And data analysts can indeed perform wonders, provided however that they have real data to work with. And this is not the case. At least not in large parts of the world.

In the end, there is no chance to have true data driven decision-making processes without true democracy, real accountability and transparency.

In the final analysis, good governance is the precondition for getting good data.