Caracas Chronicles

It’s an open secret: for the past few weeks, DolarToday (DT), the website that publishes the forbidden dollar-to-bolivar market rate, has been giving weird numbers. Everybody could tell, but nobody knew why.

See, on January 26, DT’s rate reached a historic high of 266,631 VEF/USD. It had been at 107,128 VEF/USD a month earlier — an increase of almost 150%. Mind-blowing, but not insane: we were on the fourth month of hyperinflation, the money-printing machine was doing 24/7 shifts and Maduro was creating random bonuses.

A week later, the DT dollar (DTD) was down 15.6%. I thought that’d be the classic agacharse para agarrar impulso (it’d quickly bounce back), but a couple of weeks later, the dollar lost an additional 5.8%. Some said it was the seasonal effect at full force, and the exchange rate was now correcting itself. At the end of March, with hyperinflation still compounding the bolívar’s losses, the DTD was still stuck at Bs. 200,000.

That was it: I decided to prove DolarToday was full of shit. Asdrúbal Oliveros’ mesas de dinero and AirTM’s rate were already contradicting the DTD, but I went for another approach: bitcoin.

People realized that DT was publishing below-market rates, and started shunning it (newsflash: DT doesn’t have a monopoly on information!).

The advantages of using localbitcoins.com, a popular marketplace in Venezuela, are enormous: you don’t have to waste time finding someone who wants to buy or sell bolivars, you just have to log in and choose an offer. DT claims to base their price quote on the open market, but Localbitcoins is a market you can actually analyze.

I coded a little script to interact with Localbitcoins API and download the transaction history of the Venezuelan market. Since the start of the year till the end of March, around 239,000 transactions, totalling around $40 million, were made. On average, $445,000 were moved in a day through Localbitcoins in Venezuela, sometimes reaching peak days of $800,000. This is no small market.

Using a weighted average to get the bitcoin’s price in bolivars, and Coinmarketcap’s historical closing price of bitcoin in US dollars, I created a time series for the “Dólar Bitcoin (DB).” Results show that during the first two months of 2018, the DTD and the DB went hand in hand, with the BTC dollar averaging a 1% premium over DT. This changed dramatically in March — the BTC dollar left DT in the dust. DT froze around Bs. 200,000 for no reason, while the free (bitcoin) market was passing the Bs.350,000 mark.

People realized that DT was publishing below-market rates, and started shunning it. Newsflash: DT doesn’t have a monopoly on information! The website operators must’ve noticed a drop in pageviews and finally reacted, albeit too late — it’s quickly losing its status of leading reference, as even today the DTD remains a lot lower than what people are actually willing to pay for a dollar.

But I was still curious about the behavior of DB, this new indicator I’d found, and expanded my series to include data since 2015.

In 2015Q1, DB had an average 2% premium over DT. They moved together. The DTD would move first and the DB would react a day later. The DTD was relatively stable during the first two months and rose sharply in March 2015.

Now look at 2016. Although DT was clearly the reference, during the same months the DB was priced at an average 5% discount. That year, both rates went up.

2017 was a wild ride. DTD went from 3,100 VEF/USD to 4,300, only to drop to Bs. 2,800 in a few days and climb all the way up to Bs.3,800. DB followed suit until March. When the big crash happened, DB ignored it, keeping above the Bs.3,500. DB had a discount for most of the period, but roles reversed after the crash and DB held a premium of around 28% for 21 days. I believe the market lost confidence in DT after the huge volatility of this quarter.

Data seems to validate that February is “calmer” than other months, but I don’t think it extends to March in the current conditions. The Olivera-Tanzi effect made this tax collection season relatively meaningless, and there should have been pressure on the exchange rate to depreciate more.

Should Localbitcoins displace Dolartoday?

The DB, based in localbitcoins.com, has a key advantage that makes it a better approximation of the exchange rate over DT: it’s more transparent.

I was able to access public data of transactions to do all sorts of analysis (and you can too, if you check out my code!). The daily number of transactions and volume on the site has grown significantly the last couple of years. I’ll admit that the indicator has some flaws: it’s not a very deep market, and sometimes what appears to be fake transactions can skew the graph.

But DT is much more opaque, quoting only average prices of “transactions in Caracas” or “transactions at the Colombia-Venezuela border”. How do we know they never pull numbers out of a hat? We don’t. At best, their data is faulty; at worst, they’ve been bought by market manipulators.

Nobody can keep public information contained forever. As DolarToday is increasingly recognized as fraudulent, other indicators will flourish. I’m sticking with the dólar bitcoin as the one I can trust.